Evolution of the color pattern in the microlepidopterous genus Lithocolletis

L-

n

Evolution of the Color Pattern in

the Microlepidopterous Genus

Lithocolletis

BY

ANNETTE FRANCES BRAUN

PHILADELPHIA
1914

EVOLUTION OF THE COLOR PATTERN IN THE MICROLEPI-
DOPTEROUS GENUS LITHOCOLLETIS.

By Annette Frances Braun.

CONTENTS.

I. Introduction 105

II. Methods and Observations 106

(a) Systematic Position and Characteristics of Lithocolletis 106

(6) Color Classes Represented and Structure of Scales 109

(c) Comparative Study of the Adult Markings 110

{d) Ontogenetic Development of the Color Pattern 118

1. Historical 118

2. Methods 121

3. Observations 123

4. Discussion and Conclusions 141

(e) Phylogenetic Development of the Color Pattern 149

III. General Discussion 160

IV. Summary ^ 164

I. INTRODUCTION.

The interest in a study of the origin of the many and complex forms of color
pattern in insects hes mainly in the hope that such a study may give additional
insight into the fundamental principles of evolution in general. With few
exceptions, such as the work of Tower on Leptinotarsa, researches of this nature
have been confined to the Macrolepidoptera, chiefly to the butterflies and higher
moths. The Microlepidoptera, which are less advanced from the viewpoint
of evolution, have been almost neglected. Yet, within this group, several of
the older and larger genera afford excehent and ample material for such an
investigation. To secure convincing data, a critical study, both ontogenetic
and phylogenetic, must be made of the species of a large genus, in which the
color patterns, while sufficiently diverse to indicate that a considerable degree
of evolution has taken place, are still comparatively simple.

The genus Lithocolletis fulfills to an extraordinary extent these require-
ments; hence the following investigations were undertaken with the object of
ascertaining the modes of origin of the various color patterns in this genus and
the influences which have been operative in their evolution. The study of
adult color patterns has extended over a period of several years; the observa-
tions on pupal development were made during the summer of 1910.

It is a pleasure to express to Prof. M. F. Guyer my appreciation of his
many helpful suggestions during the progress of my work.

106 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

11. METHODS AND OBSERVATIONS.
(a) Systematic Position and Characteristics of Lithocolletis.
The genus Lithocolletis belongs to the large family Tineidse of the Micro-
lepidoptera. It comprises a large number of very small moths, varying in
expanse from 4 to 12 mihimeters. The larvae are exclusively miners, usually in
leaves or very rarely in the bark of stems. The entire larval and pupal existence
with one or two exceptions is passed within the mine. The genus, as originally
estabhshed, has a remarkably uniform structure and a very characteristic
appearance, enabling one to recognize the species easily. The structural char-
acteristics which distinguish the typical and principal group are the fohowing:

,^ IP 9 ^ Face smooth; crown tufted; labial palpi por-

^-^^^^^ rected or drooping, moderately long and pointed;
maxillary palpi rudimentary; tongue of moderate
^^.^^^^^^^^^ length; antennae nearly attaining the wing length,

^^^^^^^^^=^ simple in the male, basal joint thickened and bearing

FiG.'i. Nen^iionoiLithocoiieUs. ^ P^^.^^^' ^hc forewiugs are elongate, lanceolate,

acuminate. The hindwings are about one-half the
breadth of the forewings, linear lanceolate and fringed with ciha whose length
is from four to five times the breadth of the wings. The venation is illustrated
by Fig. 1;^ forewings, lb simple, 3 absent, 4 absent, 6 absent, 7 to costa, 8 ab-
sent, 11 absent; hindwings, 3 absent, 4 absent, transverse vein absent between 2
and 5, 6 absent. The posterior tibiae bear appressed hairs.

All of the European species and the great majority of the American species
conform to the above conception of the genus. There are, however, in North
America a few species belonging to two different groups, which differ structurally
from the typical species; for these, two subgenera, Porphyrosela
and Cremastobomhycia have been erected (Braun, '08). The
former of these, Porphyrosela, differs from Lithocolletis in the
absence of vein 10 of the forewings, the absence of a pecten on
the first joint of the antennae and in the absence of hairs on the
hind tibiae.^ It thus represents a later development, phylogen- of porphyrosela
etically, in the direction usually taken in this family; that is,
a reduction in the number of veins (Fig. 2). Cremastobombycia is of a more
ancestral type than the typical Lithocolletis; it possesses an additional vein in
each wing; on the forewing, vein 4 is present and stalked with 5; on the hind-
wing, vein 6 is present and long stalked with 5 {Fig. 3) .

In the European fauna, the typical Lithocolletids are a very homogeneous
group of species both as regards larval characteristics and imaginal markings.

1 References in italics are to the text figures, those in Roman type to the figures on the
plates.

2 Since this paper was written, studies on very closely allied species have shown that these
characters are not sufficiently constant to define a genus (Meyrick, Genera Insectorum, 128me
Fascicule, p. 5, 1912). Vein 10 is probably faintly visible in large specimens of L. desmodiella.

10 g

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 107

In all, the larva is of the so-called cyhndrical type, which agrees well with the
usual type of lepidopterous larva. During the first three instars, in which the
extent of the mine is gradually increased by a loosening of one epidermis, the
larva has a characteristicahy flattened shape, with flat
head and projecting mouth-parts. After the third moult,
the larva assumes the normal cylindrical shape and the
mine becomes tent-hke by a fold in the loosened epider- ^S^TITb'^
mis. The larva then consumes the greater part of the ^S^S^^fc=^
parenchyma within the area occupied by the mine. The
pupa, except that of one species, is always formed within maMoinhydT
the mine and may or may not be enclosed in a cocoon.

The forewings of the imagoes are marked by various combinations of white
transverse fascise and streaks on a yellowish ground color. These white marks
are bordered along their inner edges by dark brown or blackish scales; some-
times, but to a less extent, dark scales may also appear along their outer edges.
The hindwings are unicolorous.

Our American species while conforming in all structural characteristics to
those of Europe, fall naturally into two groups; one of these agrees closely in
structure of larva and imaginal markings with the European type just described;
the other has what has been termed the ^^flat larva,'' and the imagoes produced
from such larvae can always be sharply distinguished by their markings from the
species of the ''cylindrical-larval group.'' The ''flat larva" retains the flat
shape, which is characteristic of the first three instars of the "cylindrical-larval
group," through the fourth, fifth and sixth instars, only assuming the cylindrical
shape and an approach to normal mouth-parts with the last moult. By some
systematists, the persistence through three additional larval instars of the curious
modification of the mouth-parts, which is in fact an adaptation for loosening
the epidermis, has been regarded as a sufficient ground for the generic separation
of this group, and the name Cameraria Chapman has been applied to it (Chap-
man, '02). The other view, which regards the modification of the mouth-parts
merely as an adaptation for a particular mode of life, and not a character indi-
cating the natural separation of that group from one with which the imagoes
agree in all structural details, has been presented in the Canadian Entomologist
for December, 1909 (Braun, '09).

The mines of the "flat-larval group" are alwa^/^s found upon the upper side
of the leaf. The increase in the extent of the mine continues through six instars,
the larva consuming only a few layers of cehs in front of it. For this reason,
the parenchyma within the mine is never entirely consumed and the mine
rendered semitransparent, as is the case in mines of the " cyhndrical-larval
group." During the seventh instar, the larva does not feed; the time is occupied
in preparation for the pupal state. Where the imago is to emerge the same
summer, most of the species spin cocoons. The cocoon is a flat, oval sheet of silk
attached around its edge to the floor of the mine, the epidermis above being

108 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

thrown into a fold. In the overwintering generation^ the larvse hibernate,
changing to pupse in spring without spinning cocoons. The characteristic which
distinguishes the imagoes of this group from the usual type is that the dark
scales appear on the outer edges of the white streaks and fasciae; if a margin is
present on an inner edge, it is much less pronounced.

The wide geographical range and the large number of species would lead to
the inference that the genus is an old one. If so, we would expect to find some
species which have halted at an early stage in their evolution and have therefore
preserved a color pattern which should bear some resemblance to the primitive
one and hence afford a clue to it. Others would have advanced farther, develop-
ment ceasing at different levels, so that the numerous differentiations produced
should constitute a chain of related species, broken no doubt by numerous gaps
but still sufficiently complete to indicate the different directions which evolution
has taken. The problem of this research is, then, to determine what is the primi-
tive color pattern of this group and knowing it to trace the paths along which
evolution has proceeded in the production of the respective patterns of the
numerous species now in existence. It is scarcely probable that the primitive
pattern is preserved among any of the earlier genera from which Lithocolletis
is descended. Such a long period of time has elapsed since the origin of Litho-
colletis that the species now belonging to ancestral genera have probably deviated
widely from the type of marking then characteristic of them. This view is
rendered more plausible by the fact that the subgenus Cremastohombycia, which
possesses structural generic characters which suggest it as the immediate ancestor
of Lithocolletis, has a type of marking differing considerably from that common
to the large and widely distributed typical group and resembling, to a marked
degree, the type of marking characteristic of some of the species of the smaller
and seemingly younger group, usually known as the '^flat-larval group. '^ If
so, we should search for the primitive color pattern among the less differentiated
species of the typical group of Lithocolletis.

If evolution takes place in several definite directions irrespective of utility,
as contended by Eimer, this group should offer an opportunity of testing this
theory, within certain limits, since many of the highly specialized differentia-
tions, indicating a high phylogenetic development and requiring a long time to
perfect, are of almost microscopic proportions and of no conceivable value in
the life of the organism. Hence it is improbable that natural selection would
have stepped in to mar any results which might have been produced through
orthogenesis.

In a genus such as this, where the color combinations are relatively simple
and the limits of the marks clearly defined and the pigmental colors belong to
the single series of yellows (or reddish yellows), browns and blacks, the problem
of tracing their differentiation both ontogenetically and phylogenetically should
be one less complicated than in higher groups, and the observations should
offer more direct testimony as to the means by which such results are obtained.

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 109

(6). Color Classes represented and Structure of Scales.

Tower (/06) has given a comprehensive survey of the hterature deahng with
the production of color in insects. Merely a brief summary of the salient facts
bearing on the subject of this research is given below, and the reader is referred
to the researches of Tower and earlier investigators for a complete statement
of what has been accomplished along this line. This author has divided the
colors of insects into three main classes, based upon their means of production.
These are chemical or pigmental, physical or structural and chemico-physical
or combination colors. Pigmental colors are further subdivided into three
groups: cuticula or dermal, hypodermal and subhypodermal colors. The
pigments through whose agency colors in the scales of Lepidoptera are produced
are largely hypodermal; though there is a possibility that certain browns and
blacks, which are very permanent under the influence of reagents, are cuticula
colors and therefore located in the walls of the scales, instead of being present
as granules within the scales. White is probably the only purely physical color
in insects, although various other colors may be produced through the agency
of physical causes acting in combination with pigments. The latter, the chemico-
physical colors, are the most widely distributed and are due to the action of
light falling upon a surface of different structural modifications over a layer of
pigment. The various lustrous, metallic and iridescent colors of scales of Lepi-
doptera are due to this cause.

In the group under consideration, the lustrous white of the streaks and fasciae
is produced entirely through structural agencies. The scales over these areas,
when tested by the addition of alcohol, cedar oil or other similar reagents — a
method originated by Dimmock ('83), — are rendered transparent and colorless,
proving that these scales are hollow and contain no pigment. The shining
golden yellows, browns, black and iridescent blues are combination colors. When
viewed by transmitted light, the scales reflecting the first three of these colors
remain essentially the same color; the iridescent blue scales appear to contain
a pale brown pigment; hence the blue is entirely due to physical causes.

The scales in the majority of the species have a remarkably uniform struc-
ture and shape, although showing great variation in size and in the proportion
of length to breadth. In any one of the typical species, L. lucidicostella or
cratcegella for example, the scales are wedge-shaped, tapering more or less rapidly
to the short stem by which they are inserted into the wing. The outer edge of
the scale is dentate, the number of teeth varying from two or three in the elon-
gate scales around the apex of the wing in the cilia to seven or eight in the broader
and shorter scales lying over the wing membrane. In some of the scales, the
teeth are approximately equal; in others, they are of uneven sizes, and a small
tooth is sometimes inserted between two larger ones. Variations in the shape
of scales over the main part of the wing are illustrated by Fig. 4- The last
row of scales projecting from the apex of the wing over the cilia is composed of

no EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

exceedingly elongate, linear scales, so that toward their bases they appear
almost as fine as the ciha themselves. Such a scale, one of the blue iridescent
scales in the cilia of L. cratcegella, is illustrated by Fig. 5. The scales forming

the apical spot, a characteristic of a considerable group
^_^ ^ '^ of species, are smaller than the average and proportion-

ately broader and are of almost uniform size. The scales
in the black patch before the apex of such species as L.
cratcegella are very similar to these.

All of the scales just described are marked with a
Fig. 4. Wing scales. serics of nearly parallel strise, there being six or eight
such ridges extending into each tooth.
In a few species, which for additional reasons to be adduced later are to be
regarded as among the farthest advanced phylogenetically, a peculiar highly
specialized type of scale has developed. The
extremel}^ brilliant luster of the white mark-
ings of L. morrisella and ostensackenella and t^ r a • , •

^ . -TIG. 5. Apical wmg scale.

P. desmodiella and of the bronzy purple me-

taUic reflections on the thorax and base of the wings of P. desmodiella is pro-
duced by these scales. These white scales, one of which is illustrated by Fig.
6, are broader than the usual type, and the sides curve outward from the base.
The outer edge of the scale, instead of being coarsely and acutely dentate as in
the ordinary scales, is obtusely dentate or, sometimes, the teeth
are entirely absent. The edge then appears scalloped, in a man-
ner similar to the edge of a pec ten. The depressions between
the strise, which appear rather broader and more widely sepa-
T. . ...... • ^^^^^ ^^^^ usual, are concave. The concavities between these

Fig. 6. White wmg ^ . . „ „ ^ , ^ .

scale oiL. morrisella. strisB appear as it tormcd by a slopmg out of the sides of the
strise. The assumption of a concave surface will explain the
brilliant and slightly opalescent luster of these scales; the various colors being
produced by essentially the same processes as they are in a soap film. These
scales contain no pigment whatever; the effects observed are due entirely to
the effect of light passing through a medium of varying thickness. It may be
noted here that in L. rohiniella, a species very closely related to L. morrisella,
this differentiation has not reached the same degree of perfection. The white
scales are, however, broader and with fewer strise than those of the more typical
species. The metahic purple scales of the base of the wings and thorax of P.
desmodiella are very similar to the white scales structurally but differ in that
they contain a brownish pigment. Their metafile iridescence may undoubtedly
be ascribed to the same agencies.

(c) Comparative Study of the Adult Markings.
The following study of the adult markings was instituted for the purpose
of discovering whether, as was suggested earfier in the paper, the primitive

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. Ill

color pattern of the genus^ or an approach to it^ still exists among the modern
species. The presence, in several species, of a number of almost straight or
slightly modified transverse bands, separated from one another by unpigmented
fascise, suggested the possibility that such a transversely banded arrangement
of pigment areas, alternating with unpigmented fascise, was the primitive color
pattern of the genus. This idea received some support from the fact that in
L. tiliacella, whose pattern will be described in detail below, the bands in the
apex of the wing are uniformly pale yellow, with no trace of darker scales. Bear-
ing in mind the generally accepted fact that, phylogenetically, pale yellow is
one of the oldest pigmental colors, some portions of the wing of L. tiliacella are
probably still in the ancestral condition. In this species, seven distinct color
areas can be made out, most of which are completely separated from one another
by white (unpigmented) areas which lie over the origins or the tips of the nervures.
As no species has been found to have a greater number of separate pigmented
areas, and most of those which do not show a banded type of marking have
fewer such areas, the conclusion was reached from the study of the adult markings
thsit the primitive color pattern consists of a series of seven transverse bands, separated
from one another hy unpigmented areas.

The verification of this conclusion rests upon the studies of pupal develop-
ment of the color pattern, which have also shown how the modifications in the
color pattern, characterizing the different groups described below, have been
brought about.

The evidence for the above conclusion as to the primitive color pattern, in
so far as it is based upon adult characters, is given below. In this relatively
brief survey of the adult markings of the numerous species comprising the genus
Lithocolletis and its subgenera, the writer has been guided in the grouping of the
species, as far as practicable at this stage of the presentation of the results of the
research, by affinities determined through the discovery of the action of certain
definite laws in bringing about changes in the shape and extent of the color
areas. Occasionally, it has been found convenient to bring in arbitrary dis-
tinctions, used to separate the species systematically. This is true in the second
group of species discussed; here the presence of a median fascia, while doubtless
a character valuable to the taxonomist, is in no way an indication that the species
are closely related to one another. Evolution has proceeded along the same
line in the different species in respect to this one characteristic alone; the differ-
entiation of the other markings has been brought about in such different ways
as to indicate the early branching of this stem of the phylogenetic tree.

In the discussion of the individual species, the characters referred to are
those developed through the action of general evolutionary tendencies. The
various levels which such development has reached before coming to a standstill,
mark the points where the different subgroups of species diverge from the main
stems. Within these subgroups, there have been numerous differentiations,
producing the actual species we now know. Such specific differences must be

112 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

minor ones, since the general plan of coloration has been already determined.
Where a group of species has branched early from the main stem, the individual
species within it will be found to be less closely related; the longer period of
time allowed for differentiation within the group has given opportunity for
greater divergence of the individuals composing it.

In general, no effort has been made to discuss the minute characters separ-
ating the species. Such characters and their processes of differentiation are
more properly discussed later.

Leaving out of consideration for the present the species of the subgenera
Porphyrosela and Cremastohomhycia, the main branch of the genus may be
divided, as has been mentioned earlier in this paper, into two natural divisions,
sufficiently distinguished from one another by the type of larva and by the
position with respect to the white fascise and streaks of the dark scales bordering
them.

In the division whose larva in the later stages is of the normal cylindrical
type, when a dark margin occurs on but one side of a white mark, it is present on
the internal edge. In the majority of the species, this margin is the only one
present; in some species, however, external margins have developed adjacent to
some or all of the streaks. The external margin is, as a rule, considerably paler
and less complete than the internal one; in a few cases, where both are equall}^
well developed, other characteristics will leave no doubt of the division to which
any species belongs. For convenience of reference, this division has been
designated Division I. The species are illustrated on Plate III.

Several well-defined groups are recognizable. Some of these include a
considerable number of species, showing close intergradations and indicating
their origin from a common ancestor at a no very remote period. Other species
seem to stand apart, showing no very clear affinities with any of the existing
species.

One group, comprising L. tiliacella (Fig. 1, PL III; Text Fig. 7), oregonensis,^
affinis,^ tritcenianella (Fig. 5), ostensackenella (Fig. 7) and mariceella (Fig. 8) as

typical forms, and fragilella (Fig. 3) and celtifoliella
Band I ^' ,. ^^[ ''' J, y\ viL (Fig. 4) as cxtremc deviations from the typical, is char-
acterized by the presence of at least two complete white

_^ ,g transverse fascise, the first crossing the wing at about

^ - -^ . , /' f. „ the basal fourth, the second near the middle.

Fig. 7. Forewmg of L. lihacella. . . ^

In tiliacella (Fig. 7), the ground color is a very pale
golden; a spot at the base of the dorsal margin is white; the two fascise are almost
straight, the first but faintly dark margined. At three-fourths of the wing length,
there are two white dark-margined spots, corresponding in position to a third fas-
cia, very narrowly separated by pale golden scales which unite with a pale band
of the ground color beyond. Then follows a fourth white fascia, separated from
the apex of the wing by a few pale golden scales which mark the position of a band

1 It was impossible to obtain figures of L. oregonensis, affinis, occitanica and alni.

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 113

of the ground color. Sometimes the apex of the wing is pure white; sometimes
the extreme tip has a few golden scales. There are thus six white spots (four
of these are the extremities of fasciae) along the dorsal margin and five on the
costal margin, separated from one another by ground color. These correspond
in position with the points of origin or with the tips of the nervures. The
positions of the white spots with respect to the nervures are as follows: the
white streak at the base extends somewhat obhquely, lying over the point of
origin of the tubular part of the upper median vein; the first fascia lies proximally
over the tip of vein 12, distally over the origin of the tubular part of the upper
median vein; the second fascia on the dorsal margin lies over the tip of vein 16;
the pair of white spots at three-fourths lie over the tips of veins 10 and 2 respec-
tively; the white fascia beyond, over 9 and 5; the last white fascia Hes over vein 7
on the costa. These white unpigmented areas are the interspaces between
bands of ground color, seven in number, of which the first two are united on
the costa.

In L. oregonensis, the same number of white interspaces are present with
the exception of the white spot at the base of the dorsum, but the fasciae are
more bent. The ground color is a deeper golden, and, whereas in tiliacella the
fasciae beyond vein 10 are represented by a few pale yellow scales only, the
ground color is distinct, and the fourth fascia is black margined. The extreme
apex of the wing, which in tiliacella showed only a faint tinge of golden, contains
a black apical spot. In affinis, tritcenianella (Fig. 5), ostensackenella (Fig. 7)
and mariceella (Fig. 8) there is a reduction in the number of white fasciae
and spots; the spots over veins 9 and 5 are wanting, and the ground color
extends uninterruptedly from vein 10 to vein 7, In ostensackenella, the white
streak enclosing the apex has become divided into a pair of opposite spots; the
base of the wing is dark brown. In affinis and tritcenianella, the white streak
before the apex does not extend entirely across the wing. In tritcenianella, the
first white fascia is situated before the tip of vein 12. As it occupies the same
position with respect to the origin of the upper median vein as the corresponding
fascia in tiliacella {Fig. 7), it would seem that the origin of this vein is the
factor determining the position of the first fascia. This conclusion is cor-
roborated by identical observations on other species.

L. fragilella (Fig. 3) and L. celtifoliella (Fig. 4) are not closely related
to the preceding members nor to one another. In both, the fasciae are
densely dusted internally with dark-tipped scales; in fragilella, the fasciae are
suffused with pale yellow; in celtifoliella, the fasciae are acutely angled, and the
oblique costal and dorsal arms reach the margins proximal to their usual position.
As will be seen later, this displacement of the extremities of the first two fasciae
proximally is of common occurrence and has been one of the principal forces in
the evolution of the species of this genus.

Desmodiella (Fig. 9), the single representative of the subgenus Porphy-
rosela, which structurally is in advance of Lithocolletis, has retained the banded

114 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

type of marking. There are two lustrous white fasciae, dark margined on both
sides, on a dark reddish brown ground color, showing a deep purple metaUic
luster at the base of the wing.

In another group of species, the median fascia only has been preserved,
either in its primitive almost straight condition or has become angulated. Most
of the species bear no close relationship to one another, and it would seem, there-
fore, that the group is merely an artificial one and that the characteristic of a
median fascia is one which appears in several divergent lines of development.
The species which may be included in this group for purposes of discussion are
the following: L. morrisella (Fig. 48), uhlerella (Fig. 49), lucetiella (Fig. 51),
symphoricarpella (Fig. 54), gemmea (Fig. 53), martiella (Fig. 52), celtisella (Fig.
56) and apicinigreUa (Figs. 55a, 556). Rohiniella (Fig. 50) shows a very clear
natural relationship with the first two mentioned species, and differs from them
only by the separation of the median fascia into a pair of opposite costal and
dorsal streaks,^ and the obliteration of the first dorsal streak by dark scales.

Morrisella (Fig. 48,), uhlerella (Fig. 49) and rohiniella (Fig. 50) are a
remarkably homogeneous triad of species; the figures will suffice to give a clear
idea of their characteristics. In lucetiella (Fig. 51), the entire basal half of the
wing is white, with the exception of a longitudinal streak from the base follow-
ing the course of the upper median vein; the apical half of the wing closely
resembles that of tiliacella (Fig. 7). In symphoricarpella (Fig. 54), the basal
half of the wing is of the golden brown ground color; the white streaks over
veins 9 and 5 are lacking. In martiella (Fig. 52) and gemmea (Fig. 53), there
is a white basal streak;^ in gemmea a white spot on the dorsal margin occupies
a position nearly corresponding to the extremity of the first fascia. In these
two species, as in morrisella. uhlerella and rohiniella, a group of small scales in
the apex forms the black apical dot. Celtisella (Fig. 56) and apicinigreUa (Figs.
55a, 556) are characterized by the presence of an acutely angled fascia and a
median white basal streak. In celtisella, the dorsal margin is also white from
the base to the fascia. ApicinigreUa is an extremely variable species; sometimes
the fascia is broken in the middle; an extreme aberrational form is represented
by Fig. 556.

Basistrigella (Fig. 57), although a somewhat isolated species, is perhaps
most closely related to celtisella (Fig. 56) and apicinigreUa (Figs, b^a, 556) of
the above group. The median pair of streaks are of equal width throughout
and each is extended at its extremity along the margin to the base.

Bataviella (Fig. 58) stands entirely apart from any of the species of
the American fauna; there are, however, several closely related species in the
European fauna. The white spot which was present at the base of the dorsal
margin in tiliacella and did not reach the costa, is here often extended entirely

1 The terms '^ costal streak" and '^ dorsal streak" are used to designate the white and usually
triangular spots situated on the costa and dorsum respectively.

2 The ''basal streak" is the white longitudinal hne beginning at the base of the dorsum and
extending along the middle of the wing for about one-third its length.

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 115

across the wing thus separating a basal patch of dark scales from the band of
similar scales beyond it. In this species^ therefore^ the number of distinct
patches or bands distinguishable along the costa is seven^ separated more or
less completely from one another by white interspaces.

The remainder of the species^ and by far the most numerous^ have been
included in one large group^ because all can be traced to a common origin and
all are manifestations of a similar tendency in development. All are charac-
terized by the presence of pairs of opposite curved, oblique costal and dorsal
streaks, widest on the margins and tapering in the middle of the wing to pointed
apices. There are a number of well-defined subgroups, due to the halting of
individuals at different stages in the general process of evolution and the differ-
entiation of a number of species from each of these.

Standing at the foot of the group and somewhat apart from the other
species, is auronitens (Fig. 10). In it, there are three dorsal streaks before the
tornus,^ but little oblique and not displaced proximally from their normal position
with respect to the nervures.

Argentinotella (Fig. 11), occitanica and leucothorax (Fig. 13) are also
characterized by the presence of three streaks on the dorsum before the
tornus. These are decidedly oblique, especially in leucothorax, in which there
is also a well-defined apical spot. In occitanica, the first, fourth and fifth costal
streaks present in argentinotella are wanting.

In fitchella (Fig. 14), there are but two dorsal streaks before the tornus;
the first is unusually large, suggesting from its position the possibihty of its
formation from two white streaks. That this has occurred and the processes
bringing it about wih be discussed later.

Salicifoliella (Fig. 15) and tremuloidiella (Fig. 16) are not closely related to
any of the large subgroup of species soon to be discussed. These two isolated
species are the result of differentiation and advancement along other lines, when
there has been a cessation, at a lower level, of the common course of evolution,
which if continued, culminates in such forms as hageni (Fig. 17) and later, in
the large subgroup of species of which cratcegella (Fig. 36) and obscuricostella
(Fig. 25) are typical representatives. Because of this, their markings will be
better understood when the laws governing their transformation have become
apparent.

Insignis (Fig. 20), hageni (Fig. 17) and arbutusella (Fig. 18) are properly
grouped together. In another variety of insignis, the white extends from the
basal streak across the wing to the dorsum, forming a quadrate white patch similar
to that of hageni. These three species are likewise differentiations from an offshoot
of the main stem, although at a higher level than the two preceding species.

We now come to the large subgroup of species characterized by two oblique

1 The three margins of the wing are the costa or upper margin, the termen or outer margin
and the dorsum or inner margin. The junction of the costa and termen is the apex of the wing;
the junction of the termen and dorsum is the tornus. In this group, where there is no defined
termen, the tornus is the place where the hue of scales in the ciha meets the dorsal margin.

116 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

streaks on the dorsum, the first usually very oblique, the second, which is placed
at the tornus, often nearly perpendicular, and four costal streaks (rarely five as
in populiella and rileyella). These are popuUella (Fig. 27), diaphanella (Fig. 28),
salicivorella (Fig. 29), deceptusella (Fig. 30), alnicolella (Fig. 32), alni, incanella
(Fig. 34), scudderella (Fig. 35), ledella (Fig. 33), malimalifoUeUa (Fig. 37), cratce-
gella (Fig. 36) and propinquinella (Fig. 38), in which the black scales in the apex
are not massed to form a circular apical dot; and carycealhella (Fig. 21), rileyella
(Fig. 22), ostrycefoliella (Fig. 23), olivceformis (Fig. 24), ohscuricostella (Fig. 25),
kearfottella (Fig. 26), obsoleta (Fig. 41), sexnotella (Fig. 39) and ceriferella (Fig.
40), in which a number of small black scales in the apex are grouped into an
almost circular, definitely outlined, apical spot. In obsoleta (Fig. 41), this spot
is very ill-defined, as are all the other markings. With the exception of the
three last mentioned species, a white median basal streak extends to about one-
third of the wing length. This streak may be margined above and sometimes
around its tip with dark scales. In some of the species, there is, on the dorsal
margin, a small white spot, often extended almost to the base of the wing and
with its outer edge about midway between the base of the wing and the first
dorsal streak. In a few cases (e. g., incanella (Fig. 34)), this spot is indistinctly
margined on its inner side by a short oblique line of darker scales. In obsoleta
(Fig. 41), sexnotella (Fig. 39) and ceriferella (Fig. 40), the entire basal portion
of the wing is without markings, the ground color extending uniformly over it.

Lastly, there is that subgroup of species in which there is but one margined
dorsal streak and that one is situated at the tornus; the basal portion of the
wing is, to a great extent, shining white, and any colored markings present are
longitudinal. These are trinotella (Fig. 47), quercialbella (Fig. 46), clemensella
(Fig. 45), lucidico Stella (Fig. 44) and argentifimbriella (Fig. 43). Phylogenetically
considered, albanotella (Fig. 42) should be included here, as representing an earlier
stage in the process resulting in the production of these five species. In this
species, a first dorsal streak is defined by an incomplete margin not reaching the
dorsum.

In Division II of the genus, illustrated on Plate IV, where the modified flat type
of larva persists through six larval instars, the chief aggregation of dark scales is
on the outer side of the white streaks and fasciae. Where an additional margin
occurs on the inner side of the white markings, the outer one has broadened con-
siderably and is often prolonged at the angle of a fascia as a dense aggregation of
dark tipped scales. There seems to be less tendency in this group toward the con-
centration of dark scales in the apex over a definitely outlined area ; the tendency
is rather for all the scales to become dark tipped, producing an evenly dusted
effect. As in Division I, several groups of species are recognizable.

In one group, there is a tendency to preserve, either straight or angulated,
two white fasciae, the first crossing the wing at the basal fourth, the second at
the middle. In obstrictella (Fig. 64), these fascise are almost perpendicular
and straight; the pair of white streaks beyond often form a third fascia. In

EVOLUTION OF COLOE PATTERN IN LITHOCOLLETIS. 117

tubiferella (Fig. 65), aceriella (Fig. 66) and hamameliella (Fig. 67), these
fasciae are straight and but httle obUque. In tubiferella (Fig. 65), they are the
only white markings, except sometimes a very small white dot near the apex.
In aceriella (Fig. 66) and hamameliella (Fig. 67), there are, in addition, an
obhque white streak at the base of the dorsum, a longer obhque dorsal streak at
the tornus and a small white streak opposite it on the costa. Ostryarella (Fig.
6S) , corylisella (Fig. 69), CBSculisella (Fig. 70) and guttifinitella (Fig. 71) constitute
a closely related series, separated from one another by small differentiations.
In these species, the two fascise are bent outwardly below the costa, scarcely
enough to be called angulated. In guttifinitella, the pale streak at the base of
the dorsum is almost or sometimes entirely wanting. In lentella (Fig. 73) and
carymjoliella (Fig. 72), the fascise are distinctly angulated.

In cincinnatiella (Fig. 75), macrocarpella (Fig. 74), saccharella (Fig. 76),
hamadryadella (Fig. 77), umbellularice (Fig. 78) and agrifoliella (Fig. 79), at least
one of the two white fascise has become acutely angulated. In hamadryadella
(Fig. 77) and umbellularice (Fig. 78) , the white fascia at the base is preserved to the
costa, although in umbellularice, it is almost separated into two parts. In agrifoliella
(Fig. 79), the position of this fascia is indicated by two patches of dark scales which
would form its external margin. In hamadryadella (Fig. 77), very complete
internal margins to the white fascise have developed; the ground color has,
however, shrunk away toward the base, so that the line of dark scales extends
through the middle of a broad white band. In cincinnatiella (Fig. 75) and
macrocarpella (Fig. 74), the two fascise are uniformly angulated; the white patch
at the base does not reach the costa. The costal arm of the first fascia of
saccharella (Fig. 76) is wanting, its dorsal arm is very obhque and confluent with
the oblique basal streak.

A group of species compiking fl.etcherella (Fig. 80), arcuella (Fig. 81), betuli-
vora (Fig. 82), australisella (Fig. 83), bethunella (Fig. 84), chamber sella (Fig. Sb),
-cervina (Fig. 86), platanoidiella (Fig. 87), castaneceella (Fig. 88) said fasciella (Fig.
89) retains but the single median fascia, sometimes almost divided at its angle.
A pair of opposite oblique streaks represent the former complete fascia at the
basal fourth of the wing in the five first mentioned species of this group; in the
remaining five, the costal streak only is present. In castaneceella (Fig. 88) and
fasciella (Fig. 89), there has been an extensive progressive diminution in the white
markings; the median fascia is the only white marking remaining in fasciella.

In gaultheriella (Fig. 91) and nemoris (Fig. 90), there is no median fascia;
instead of it a pair of costal and dorsal streaks. In nemoris, the first pair of
streaks meet, forming a fascia; in gaultheriella, they are separate.

Lastly, there is that group of four species, in which a white longitudinal
streak extends from the base along the dorsal margin, reaching to or beyond the
middle. These are mediodorsella (Fig. 92), quercivorella (Fig. 93), ulmella (Fig.
95) and conglomeratella (Fig. 94). In the two former, this streak stops abruptly
at the middle of the dorsal margin; in mediodorsella (Fig. 92), it connects with

118 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

the dorsal arm of an angulated fascia. In ulmella (Fig. 95) and conglomeratella
(Fig. 94), the white streak extends almost to the tornus, where it connects with
a dorsal streak running obliquely into the apex.

A rather noteworthy feature common to both divisions of the genus is this :
the third white fascia or the third pair of streaks, whose extremities are placed
over the tips of veins 2 and 10, are never displaced from their normal position,
and only in rare instances are they much oblique. These instances are almost
entirely confined to Division 11. In such case, the impression of length and
obliquity is gained through the fact that the more apical portion of this streak
is homologous with the white streak over vein 5. The greatest changes have
taken place in the positions of the first two pair of streaks, resulting in their
displacement toward the base.

In either of the two divisions of the genus, the greatest number of more or
less distinct bands of ground color has been seven. The first of these crosses the
base of the wing and is usually confluent with the second in the costal half of
the wing. The position of the white unpigmented areas separating these bands
of ground color appears to be determined by the points of origin or by the tips
of the longitudinal nervures. The question to be decided is whether such a
banded arrangement of pigment areas represents the primitive type of colora-
tion, and if so, how has evolution acted in bringing into existence the present
great complex of species.

A group of species (Figs. 59a, 596, 60, 61, 62, 63) more primitive in
generic structure, namely, Cremastobomhycia, shows a type of coloration strik-
ingly similar to that of some of the species of the ^^flat-larval group." An
examination of this group discloses the fact that nowhere are there more than
four costal white streaks as compared with five or six in some of the species
of the more recent branches of the genus. Apparently, then, because of the
fewer separate bands of ground color present in species of this group, the primitive
type of marking is not to be sought for among the species of the ancestral group.

The observations that have been made point to the conclusion that the
primitive type of coloration is made up of seven distinct transverse bands. The
results of the investigations upon the development of color in the pupal wings,
now to be presented, confirm this view and, in addition, furnish evidence of the
means by which transformation of the bands into the various types of color
pattern has been accomplished.

(d) Ontogenetic Development of the Color Pattern.

1. Historical.
Tower ('06) has summarized the general results of researches upon the de-
velopment of colors in pupal wings as follows :

^^The studies upon the origin of color in the wings of Lepidoptera have
shown that there is a regular order of development, not only of colors but also
of the areas in which spots and stripes appear. The wings, which are colorless

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 119

at first, become opaque, yellowish or light drab, beginning first near the base of
the wing and spreading distalward. Soon spots, stripes and adult markings
appear, at first proximad, then more and more distad, then between the nervules
and, last of all, upon the nervules. First, the purely pigmental or chemical
colors develop, and these are followed by the chemico-physical.'^

The principal researches have been those by Schaffer (^89), Van Bemmelen
('89), Urech ('91), Haase ('93), Mayer ('96) and most recently by von Linden
('98, '02). The principal work of the first three authors was upon Vanessa and
Pyrameis. Van Bemmelen ('89) has shown that the first color to appear in
Pyrameis cardid is a pale brownish yellow which deepens to the reddish brown
ground color; upon this ground color, the black spots appear. The development
of the color in Vanessa urticce follows very much the same sequence, as pointed
out by both Van Bemmelen and Urech. Urech ('91) has shown that in Vanessa
10, the adult spots and colors appear directly upon the primitive white of the
pupal wing; that is, the yellow appears over a restricted area of the wing;
following this, a reddish tinge appears, which, over another portion of the wing,
gradually deepens to the reddish brown ground color. Later, the adult black
spots develop directly upon the primitive white wing; some at least of the white
spots of the adult represent the primitive white of the pupal wing. In Van
Bemmelen's opinion, a considerable amount of modification takes place in the
wings after the first appearance of color, so that the adult design may differ
considerably from that first laid down. Urech's ('91) view is directly opposed
to this; according to him, the different color areas are definitely laid down from
the beginning and are therefore older phylogenetically than the particular colors
of the adult wing. He therefore regards color as the important factor in deter-
mining phylogenetic position— ''die Farbe ist das Primare, die Zeichnung das
Secundare."

Haase ('93) worked upon several species of Papilio and found that the
wings are transparent or colorless, later becoming whitish; this white then gives
way to a yellowish ground color upon which the adult colors begin to appear,
undergoing considerable development before reaching their definitive adult
condition.

Mayer ('96) traces the development of the scales from modified hypodermis
cells, the ''formative cells of the scales," from their first appearance as blunt
outgrowths from these cehs until they are fully formed and pigmented. In
addition to confirming the conclusions of previous investigators as to the sequence
in the appearance of colors, Mayer has shown that "the transparent condition
of the wings corresponds to the period before the scales are formed and to the
time when they are still completely fuh of protoplasm. The white condition is
caused by the withdrawal of the protoplasm from the scales, leaving them as
little hollow bags filled with air. In this condition, they diffract the light and
appear pure white.

"After the protoplasm has completely withdrawn from the scales, the

120 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

'blood' or hsemolymph of the pupa enters them; and soon after this, the wing
becomes of a uniform dull yellow or light drab color. This color is due to the
fact that soon after the hsemolymph has entered the scales, it changes to a dull
ochre-yellow and finally to a drab color. The same change takes place in hsemo-
lymph which has been removed from the pupa and exposed to the air. The
mature colors are clue to chemical changes in the hsemolymph itself. They
first appear in places between the nervures, never upon the nervures themselves.
The last places to acquire the mature coloration are the outer and costal edges
of the wings and the nervures. . . .

''Dull ochre-yellow and drabs are, phylogenetically speaking, the oldest
pigmental colors in the Lepidoptera, for these are the colors that are assumed by
the haemolymph upon mere exposure to air. The more brilhant pigmental colors,
such as bright yellows, reds, greens, etc., are derived by more complex chemical
processes. We find that dull ochre-yellows and drabs are at the present day the
prevalent colors among the less differentiated nocturnal moths. The diurnal
forms of Lepidoptera have almost a monopoly of the brilhant colorations, but
even in these diurnal forms, one finds that duh yehow or drab colors are stih
quite common upon those parts of their wings that are hidden from view.''

The most recent researches upon the subject of development of color in
the pupal wings of Lepidoptera are those of von Linden ('98, '02). The results
of her investigations are embodied in two papers, the first being practically in-
cluded in the second. Her observations lead to the conclusion that the adult
color pattern develops from a number of separate transverse bands which dispose
themselves upon a uniform ground color in a manner characteristic of each
species. A distinction is thus made between ground color and that forming the
color pattern. These transverse bands (longitudinal according to Elmer's view)
occupy definite positions upon the wing determined by the course of the nervures
and trachese. This is the condition found among Papilios, Vanessas, Sphingids,
the higher Bombycids and even among some of the Geometrids. These larger
bands are formed by the broadening out and fusion of narrower bands ("bande-
lettes primaires") preserved only among more primitive forms. On the one
hand, the bands may fuse to form a uniform color; on the other, they may break
up into a series of spots, which become smaller as the ground color encroaches
upon them. The hindwing is arrested at a point less advanced than the fore-
wing; this, combined with the greater tendency toward fusion of the bands
which is in fact due to the modified form of the hindwing, creates the erroneous
impression that the hindwing is more advanced than the forewing.

The author claims that her results support the laws which Eimer has promul-
gated for species differentiation based upon a study of adult coloration :

"/Z est done evident que les his que Eimer a trouvees en etudiant la phylogenese
du dessin des Lepidopteres peuvent etre admises point pour point pour Vontogenese.

"Mes reserches preuvent done que le dessin des Papihons ne parait pas
soudainement, . . . que les phases que ce dessin parcourt pendant son developpe-

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 121

ment ontogenetique, sont analogues aux phases que le dessin de Tespece du
Papillon a dti parcourir pendant son developpement phylogenetique, que le loi
qui gouverne la biogenese est aussi confirmee par ce procede, qui s'accomplit
pendant la metamorphose de Tinsecte^ et qu'en fin les theories de Eimer sont
solidement fondees, quand il dit que chaque animal ne pent varier que dans peu
de directions determinees par les influences du dehors et par sa propre consti-
tution, et que Fespece se forme par genepistase/^

Recapitulation is, however, only partial and confined to primitive forms:
^^Les phenomenes de la phylogenese se produisent le plus clairement dans
le developpement du dessin chez les formes les plus primitives. Chez les groupes
les plus advances, ces phenomenes sont masques par ^intervention dela tachy-
genese qui montre sa plus grande influence dans le developpement des ailes
inferieures.''

With regard to the succession of colors upon the wing, the same author
finds that it is only in the primitive forms that the scales forming the dark mark-
ings pass through the intermediate shades of yellow to reach their final gray or
black condition; in the higher forms, they remain uncolored until the scales
forming the ground color reach their full development, later passing directly
from the colorless stage into gray or black. The color appears first at the tip
of the scales, spreading gradually to the base. Scales which are destined to
acquire their pigment later are not fully formed when the scales of the ground
color attain the adult form and color. Therefore, color may be taken as a
measure of the degree of development of a scale.

2. Methods.

In the study of the development of color in the wings of Lithocolletis , speci-
mens of the summer generations were used for observation. No overwintering
pupse have been observed. In the summer generations, the average duration
of the pupal state is about a week; it may be as short as four or five days.

The chrysalids were removed from the mines shortly after pupation and
kept under careful observation. For some time, no change in the brown color
of the pupa is to be observed, then two faintly darker brown spots, corresponding
in position with the eyes of the imago, become visible through the pupal envelope.
These spots rapidly turn black, and very shortly after, the wings, which up to
this time have appeared transparent through their thin chitinous envelope,
begin to lose their transparency. This marks the beginning of the white stage.
A day or two later, the yellow colors begin to appear upon the wing. The chief
difficulty in the study lies in securing a suitable series of these earliest stages,
as the colors are very faint at this period, and numerous dissections are often
required before the necessary material can be obtained. In later stages, the
markings are visible through the pupal envelope, and a choice of specimens can
be made with relative accuracy.

The dissections were made in .6 per cent NaCl solution under a simple

122 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

dissecting microscope of moderate power by means of fine needles set in match-
sticks. One needle is inserted through the abdomen and held rigid; another
through the middle of the upper side of the thorax. The second needle is now
drawn forward, tearing apart the pupal envelope and carrying with it the anterior
part of the thorax with wings attached. The structure of the pupa with separate
envelopes for wings and legs in the lower Microlepidoptera renders this method
of procedure highly successful. The thorax may now be torn apart, allowing
the wings to float out flat. With a little experience, it is possible to remove
the wings in this manner without the displacement of either scales or cilia. The
wings thus removed from the pupa vary in length from L5 to 2 mm., being,
therefore, about one-half the length of the imaginal wings.

The wings were examined with a simple lens to determine the general extent
of the color areas. For the more accurate observation of the positions and limits
of pigmented areas and individual scales pigmented, I have used a compound
miscroscope, 16 mm. objective, 4x eyepiece, giving a magnification of 60 diam-
eters. The observations were made almost entirely by reflected light and
under a cover glass. In the early stages, where the colors are very faint, I have
found black paper, as suggested by von Linden, very useful for a background.
Somewhat later, a pure white background is advantageous in the detection of
the first appearance of grayish tints.

All of the wings were examined in .6 per cent NaCl solution, shortly after
removal from the pupa. Observations made thus are vastly more satisfactory
than when permanent mounts are employed. A number of the wings were,
however, transferred to absolute alcohol between pieces of glass to prevent
curling, and after about an hour embedded in Venetian turpentine made as
follows :

Commercial Venice turpentine is mixed in a tall cylinder glass with an
equal volume of 95 per cent alcohol. The mixture is aflowed to stand in a warm
place for three or four weeks and then decanted.

The mounting medium thus prepared is thin enough to flow easily and not
displace the deUcate cilia. However, the attendant darkening of the colors is a
disadvantage in comparing with adult specimens.

I have found it possible to preserve all except the earliest stages without
the use of a mounting fluid, if the slide containing the wing under a cover glass is
immersed in absolute alcohol for a short period. Later the cover glass may be
closely appressed and fastened down around its edges.

Both forewings in each specimen were examined; the two sets of observa-
tions thus serving as a check on one another. In the following descriptions,
reference is made to the forewings only. The hindwings are practically uni-
colorous, usually paler than the forewings, and reach their final development
somewhat in advance of the forewings. Their development yields no new
facts, merely confirming the results of earlier investigators upon the sequence
of colors.

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 123

3. Observations.

Two factors were taken into consideration in determining the choice of species
to be studied, namely, the general type of adult marking and the presence or
absence of certain characters such as an apical spot or dark marginings adjacent
to white streaks or fascise. The object of the research is twofold: (1) to deter-
mine, if possible, the primitive or fundamental color pattern and its mode of
transformation into the adult pattern and (2) to determine the relative time of
appearance and manner of development of the various dark markings, the
characters referred to above.

The evidence furnished by the development of color in the pupal wings
taken in conjunction with a comparative study of the adult markings has con-
vinced me that the primitive color pattern is composed of a series of uniformly
colored pale yellow ban ds, seven in number, placed transversely on the wing and
separated from one another by white fasciae (unpigmented areas) (Fig. 8) . These
bands are typically straight and, with the exception of

the first, cross the wing almost perpendicular to the mar- j, ^^ ^ v vi

gms. The bands, when present m their primitive con- ■ ' '

dition, have certain defined positions. Their hypothetical
typical positions are illustrated by Fig. 8. In what fol- '" ^

lows, it will be noticed that there is a definite relation _ ^^ ,.,...

. . Fig. 8. Hypothetical primitive

01 bands to veins. Band I occupies the extreme base color pattern of the forewing of
of the wing, beginning at the base of the dorsum and Lithocoiietis.
broadening on the costa; the white streak which sepa-
rates it from Band II passes obliquely from the base of the dorsum toward
the costa, lying over the point of origin of the tubular portion of the lower
median vein. Band II crosses the wing on the basal side of the tip of vein
12; the position of the white fascia which separates it from Band III seems
to be determined rather by the point of origin of the tubular portion of the
upper median vein than by the tip of vein 12, as this vein often extends con-
siderably beyond the origin of the upper median vein without causing a cor-
responding broadening distally of the second band. It sometimes happens
that vein 12 is unusually short, and then the white fascia lies over its tip and
over the origin of the upper median vein. On the costa, the inner border of
Band III hes just beyond the origin of the upper median vein, and on the dorsum,
the outer border extends almost to the tip of vein 16. Band IV crosses beyond
the tip of vein 16 and hes over the bases of veins 10 and 2, its outer border being
just within the tips of veins 10 and 2. Band V on the costa is placed between
veins 10 and 9, on the termen between veins 2 and 5 and lies over the transverse
vein. Band VI is placed between veins 9 and 7 on the costa and beyond vein 5
on the termen. Band VII crosses the apex of the wing. Thus it will be seen
that the position of the bands is largely determined by the position of the longi-
tudinal nervures, the tips of the veins marking the extremities of the unpig-
mented fasciae separating the bands. In no case in the adult has this pattern

124 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

been preserved in its original primitive condition as seven distinct transverse
bands; fusions either partial or entire have taken place between two or more
adjacent bands. Displacements have also occurred^ due to a shrinking away
of color from one side of a white interspace and its extension over the white on
the other side. Finally^ some of the bands may have shrunk away almost
entirely.

The following is a list of the species in which the development of the pattern
has been traced from the first appearance of color to the adult markings :

Lithocolletis tiliacella Cham.
Lithocolletis tritcenianella Cham.

Lithocolletis cratcegella Clem.
Lithocolletis ostrycefoliella Clem.
Lithocolletis lucidicostella Clem.
Lithocolletis morrisella Fitch
Lithocolletis hamadryadella Clem.
Lithocolletis hethunella Cham.
Lithocolletis ulmella Cham.
Lithocolletis cesculisella Cham.
Cremastobombycia ignota F. and B.

Species in which two or more of the bands
have been preserved almost in their
primitive condition as respects position
and shape.

Species in which changes in shape or
position or both and more or less com-
plete fusions of some of the bands have
taken place and certain differentiations
and specializations have appeared, such
as the development of black scales at
places other than those directly con-
tiguous to the white interspaces be-
tween the bands.

In the following species, the development of the pattern was observed only
in part for the purpose of comparing the development of certain characters in
these species with that of similar characters in related species, or with a view of
determining the relative time of appearance of certain characters:
Lithocolletis hageni F. and B.
Lithocolletis robiniella Clem.
Lithocolletis carycefoliella Clem.
Cremastobombycia solidaginis F. and B.

In all of the species studied, the wings are at first transparent and glassy
in appearance, corresponding to the period, as observed by Mayer (^96), before
the protoplasm has been retracted from the scales. Following this is the so-
called white stage, in which the wings appear by reflected light almost pure white
and by transmitted light pale bullish, due to the hsemolymph contained within
them. The scales are at this period colorless and transparent, containing only
air. The white stage lasts from one to two days, giving way to a pale yellow
over certain defined areas, as the hsemolymph enters the scales and the pigment
begins to form. This yellow color may be preserved throughout pupal develop-
ment almost as originally laid down, or it may go through a process of transfor-
mation before reaching the adult condition. In either event, it gradually deepens
in tint and constitutes the ground color upon which the dark streaks (margins)

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 125

and spots are to appear. The development of the pattern takes place very
rapidly during the last part of the pupal state. The time intervening between
the first appearance of a tinge of yellow on the wings and the emergence of the
imago is largely dependent upon temperature and is often less than twenty-four
hours. Development takes place more rapidly in day time than at night and
is strikingly retarded during cloudy weather.

Lithocolletis tiliacella Cham.

In the first specimen examined, the wings were removed from the pupa at
a very early stage of development. The yellow color is scarcely differentiated
from the clear buffish white wing {Fig. 9) . The only bands ^ jtr jy

which can be distinguished are II, III and IV. Band II

crosses the wing just before the tip of vein 12, thus occu- ' j

pying its primitive and typical position. It is the palest

and least defined of any of the three bands present. Its Fig. 9. Early stage in the
outer edge is definitely defined, but its inner edge fades d^^^^^P^^^t ^f f^^ ^^ t^e

^ ^ ^ ^ ^ wmg 01 L. hhaceUa.

mto the clear basal part of the wing, the color being ex-
tended farthest toward the base just within the costal and dorsal margins.
Band III has also preserved its primitive position and form. It is somewhat
darker than II, but uniformly colored throughout its breadth, the outer edge
not darkened. Its width is scarcely greater than that of the clear band pre-
m jv V ceding it. Band IV is the most deeply colored as well

JI \ I ; VI YU

^ ' . -' i as the broadest of the three bands; it is of the same

i ' breadth as in the adult. Even at this early stage it is

prolonged outwardly almost to the end of the cell. The

Fig. 10. Later stage in entire apical part of the wing and cilia are of the

the So^nklr'" " ^^^^^ ^hi^i^h ^^1^^^ ^he^^ '^^i^g ^o indication whatever of

bands.
The wing in the next stage examined shows considerable advance over the
stage just described {Fig. 10). There has been a decided deepening of color,
and the adult pattern has been laid down, with the possible exception of Bands
VI and VII, which have not as yet been clearly differentiated as bands although
there is a faint yellow tinge in a few of the scales in the apical portion of the wing.
These bands are usually indistinct and sometimes absent even in the adult.
Band I is united with Band II along the costa; this condition was apparently
not produced by the independent origin of Band I and its later fusion with II
but by a uniform and gradual deepening of the wing below the costa. The
outer edge of Band II is now somewhat bent outward in the middle, and the
scales in the middle of the outer edge are beginning to deepen in color, thus
foreshadowing the appearance of the dark margin, but there is as yet no hrown
pigment in the scales. Band III is also somewhat bent outwardly as in the
adult, and its dark margin is present on the costa and dorsum only, the costal
margining being much less distinct than the dorsal; neither extends more than a

126 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

short distance onto the wing proper. The middle of Band IV is faintly con-
nected with a pale band beyond, which represents Band V. Its margins have
reached more nearly their adult condition than any other of the wing markings.
The margin on the dorsal half is almost as greatly developed as in the adult;
that on the costal half is represented only by a few black scales.

The next stage studied shows no further change in the shape or extent of
an}^ of the bands, the changes which have taken place being confined to continu-
ation of processes already begun, namely, the deepening of the outer edge of
Band II, thus forming the internal margin of the first white fascia in the adult,
the approximation of the costal and dorsal parts of the margin of Band III, so
that they are now but little separated, and their darkening to the adult color,
the advance of the margins of Band IV to their adult condition, the deepening
of Band V so that it now appears quite distinct and as in the adult.

This is the only species studied which shows any distinction between the
bands in regard to their relative time of appearance. Bands II, III and IV are
the first to appear, and in the youngest specimen examined, Band IV was de-
cidedly in advance of the other two as shown by its deeper color and by its
extension outwardly in the middle. It is, therefore, probably safe to assume that
Band IV is in this species, the earliest to appear ontogenetically. As will be
shown later, L. tiliacella has probably in many respects conserved more closely
than any other of the now existing species the primitive type of color pattern,
and hence the sequence observed in the appearance of the bands in the ontogeny
may also be true for the phylogeny.

Lithocolletis tritaenianella Cham.

At the earliest stage at which it is possible to discern any indication of
marking, five very pale yellow bands are visible upon the wing {Fig. 11). These

are uniformly colored and straight and are placed trans-

^^' versely upon the wing, showing no signs of fusion with

J one another. The first of these is, in fact, I + II, as

shown by the faintly paler streak extending into it from

Fig. 11. Early stage in the the basc of the dorsum. The sccond. Band III, occu-

deveiopment of color in the pj^g j^g ^g^^j positiou, but in this casc the position of

wing of L. iriioenianella. ,. p. . .^ tx* inij

the white lascia separating it from II is wholly deter-
mined by the point of origin of the upper median vein, as vein 12 extends for a con-
siderable distance into Band III. The third, Band IV, is normal in position and
shape. That the fourth band is in reality composed of Bands V and VI is shown
by its origin just beyond the tips of veins 10 and 2 and by its extension distally
over the tips of 9 and 5 which in the primitive color pattern mark the position
of the white fascia separating Bands V and VI. Band VII crosses the apex of
the wing. The white interspaces between Bands II and III and between III
and IV are of about the same width as in the adult, but that between IV and
V + VI is considerably wider and shows no indication of the angulation which

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 127

is present in the adult; that is, Band IV has not yet begun to extend outwardly
through the cell. The space separating V + VI from VII is at this stage about
one-half the width of V + VI and extends entirely across the wing, thus differing
strikingly from the condition found in the adult.

The development for some time consists merely in a uniform darkening of
the ground color as first laid down without any change in the extent of the bands.
At a somewhat later stage, simultaneously with the darkening of the ground
color into the adult shade, changes in shape and more complete fusions of some
of the bands occur. As soon as the adult color has been attained, the dark mar-
gins begin to develop. A specimen examined at this period shows that while
no change has taken place in the form or extent of Bands I + II and III, Band
IV has become outwardly angulated as in the adult, thus narrowing, especially
in the middle of the wing, the white fascia separating it from V + VI. On the
dorsal margin. Bands V, VI and VII are completely fused, the small white costal
streak before the apex alone remaining to mark the position of the white fascia
between V + VI and VII. The scales along the outer edges of Bands II and
III are beginning to show faintly brownish tips, this color being about evenly
distributed along the entire edge of Band II; in Band III, the color extends
farther toward the bases of the scales in the middle of the wing. The scales of
the dorsal half of the outer edge of Band IV are dark reddish brown, almost as
in the adult, but on the costal half of the wing, there is only a very slight deepen-
ing of the yellow color, the margin here being considerably paler than the margins
of Bands II and III. There is a very slight deepening of scales on the outer
border of Band V + VI, adjacent to the white streak before the apex. The
dusting in the apex is dense, the color being darker and more blackish than the
outer margin of Band IV, and extends from the white streak well around the
apex and along the dorsal margin almost to vein 5. This dusting is considerably
greater in extent even in this stage than that of the overwintering forms in any
stage.

Lithocolletis crataegella Clem.

This species belongs to that large group in which the imago is characterized
by the presence of pairs of opposite white costal and dorsal streaks; that is, all
of the original primitive bands have fused at least along the middle of the wing.
L. cratcegella (Fig. 36, PI. Ill) possesses in addition a margined basal streak, ex-
ternal as well as internal margins to the white streaks (i. e., the dark scales
developed on both sides of the primitive bands and not only on the outer side, as
in the two preceding species), a streak of dark scales along the middle of the
wing, a patch of black scales in the apex and the blue iridescence of the scales in
the cilia. The study of the development of the wings in this form should there-
fore throw light upon the origin and mode of development of the above characters
and indicate their relative time of appearance.

In the earliest stage examined (Fig. 12), at a period when the markings are

128 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

pale yellowish and but little differentiated from the colorless wing, the pattern
strikingly resembles that of the adult insect. The white basal streak is distinctly
outlined, beginning at the base of the dorsum, where it seems to be homologous
j^^jjfj. IV. V yj: with the white streak separating Bands I and II, be-

yond this lying over a trachea which extends through
' / the cell just above the middle; its tip seems to coincide
with some portion of the white fascia separating Bands
Fig. 12. Early stage in the II and III. The white patch near the base of the dorsum
development of color in the wing • ^ identical with that in the adult and with the apex of the

01 L. cratcegeUa.

basal streak seems to constitute the last vestige of the
white fascia separating Bands II and III. On the dorsal margin, the outer edge
of Band III has shrunk away toward the base, and at the same time, there has
been a compensating extension of Band IV toward the base, resulting in a dis-
placement of the white streak, so that instead of being over vein 16, it is basal
to it, and vein Ih reaches the margin well within Band IV. The first pair of
streaks are already separated by a narrow line, which is not nearly as broad
internally as in the imago. At this stage, the costal white streak is not much
narrower on the margin than the dorsal. Bands V, VI and VII are situated
normally, except that V on the dorsum extends a little basally over the tip of
vein 2. These bands, while quite distinct on the margins and in the cilia, seem
to fade away in the middle of the wing. The entire middle of the wing, from the
tip of the first dorsal white streak almost to the extreme tip of the wing, remains
colorless, so that it is not possible to distinguish the bands from the interspaces
separating them. A comparison with the adult shows that this area coincides
with that to be occupied by the dark streak of scales through the middle and
by the apical patch of black scales.

In a specimen somewhat more advanced, there has been a slight deepening
of color, but by no means, even in the golden color, has the adult shade been
attained. The color is still yellow, with no tinge of brown, unless the slight
darkening of the dorsal side of Band IV, especially toward the second white
streak, may be regarded as a foreshadowing of that color. The middle of the
wing from the tip of the first dorsal streak almost to the apex (except for the
grayish black scales to be noted below) still retains its whitish appearance.
Just before the apex and beginning just under the tip of vein 7, and extending
for a short distance along the underside of 7, are a number of grayish black
scales. These appeared first beneath the tip of 7 and were gradually extended
toward the base. The patch is broadest toward the apex, where it reaches al-
most across the wing. The inner edges of the first pair of streaks are more curved
than in the preceding specimen as a result of the broadening internally of the
yellow line separating them. The scales attached around the apex are yellow
and extend in an unbroken line to the second dorsal streak, that is, Bands V,
VI and VII have united along the termen. However, the scales whose apices
project beyond these and which are destined to constitute the iridescent blue
line in the cilia still remain pure white.

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 129

A specimen at a somewhat later stage {Fig. 13) shows a httle more of the
orange brown tint, but the color is nowhere as deep as in the adult. The formerly
colorless area in the middle of the wing is still distinctly set apart from the rest
of the wing but has become narrower and is beginning to acquire a faintly yel-
lowish tinge. The white streaks are now definitely blocked out due to the
curving in of the yellow color; they are, however, still connected with the whitish
middle portion of the wing. The patch of blackish scales near the apex has
increased in extent and depth of color until its identity with the patch of hlack
scales in the apex of the adult wing is clearly established. It has now reached
its adult condition, and no further change takes place. ^j^jf^ jv y ^,j

The early origin and differentiation of this patch of ^^T-^-'
scales shows clearly that it is a character quite sepa- ' ^

rate and distinct from the dark line of scales in the

adult extending from it toward the base of the wing. Fig. 13. Later stage in the
The scales around the apex, yellow in the earher stages, development ^f color in the wing

*^ ^ of L. cratoegella.

are now tipped with pale brown, but the outermost row

of scales are still whitish as before. There is some deepening of the ground
color just before the first and second pairs of streaks; the second dorsal streak,
however, is the only streak which can be said to have acquired a definite mar-
gin, and even this is pale brownish. The basal streak remains entirely un-
margined.

In a specimen at a little later stage than that just described, internal margins,
though very pale, have appeared along all of the white streaks; that of the second
dorsal still continues to be in advance of the others. The margin of the first
costal streak has deepened somewhat, and there is a faint suggestion of gray
around its tip. A similar pale gray shade extends from the tip of the first dorsal
streak through the middle of the wing and around the tip of the second dorsal
streak. The ground color along the upper edge of the basal streak is somewhat
darkened, but there is no margin.

In a wing at a considerably later period of development, the ground color
approaches the adult color over those areas where there are no brown scales.
The internal margins of the costal and dorsal streaks, as well as the upper margin
of the basal streak, are dark brown; the external margins of these streaks and
the margin around the tip of the basal streak while present are not as dark as
in the adult and show a decidedly grayish tinge. This same gray color is pro-
longed from the tip of the first dorsal streak along the middle of the wing in the
area formerly colorless but is much paler than in the adult. This difference from
the adult color is especially noticeable above the patch of black scales, where
in the imago there is a decided deepening of color, almost blending with the
black patch. There is but very httle deepening of color between the first and
second dorsal streaks, and the margin of the second dorsal streak is neither as
deep nor as wide as in the adult. The darkening of scales in the cilia started in
earher stages stih continues. There is, however, no change from the colorless
condition of the scales which are to be deep iridescent blue.

^^ IT/

130 EVOLUTION OF COLOR PATTERN IN LITIIOCOLLETIS.

The adult coloration appears to be attained several hours previous to
emergence. At this time the markings are plainly visible through the pupal
envelope. The bluish iridescence in the cilia is now well developed. Obviously
this is one of the last characters to appear ontogenetically.

Lithocolletis ostryaefoliella Clem.

This species (Fig. 23, PL III) possesses an apical dot in its typical condition,
composed of small black or blackish brown scales arranged in an almost circular
shape over the tip of the wing.

In the youngest specimen examined, the wings were dissected out at a period
when they were losing their whitish appearance and becoming faintly banded.
i,ii,m j^ V ,., The last four bands are straight and placed transversely

on the wing, leaving colorless bands between them. The

basal portion of the wing is exceedingly pale, and the

ground color is scarcely discernible on the otherwise

Fig. 14. Early stage in the colorlcss wiug. The Configuration of the color areas is

development of color in the wing ^^^^ ^^^ ^^^^ ^^ -^ ^^^ ^^^^^ there being the Same

of L. ostrycefokeUa. ' ^

outward angulation of Band III, internal to the first
pair of white streaks. A stripe along the middle of the wing and a white spot
on the dorsal margin remain colorless and mark the position of the basal streak
and the white dorsal spot in the adult.

At a slightly later stage, the markings just referred to have become a little
more distinct; the color is still a very pale straw {Fig. IJj). The extreme base of
the wing, while very pale, is still sufficiently contrasted with the unpigmented
areas to show the outline of the basal streak defined for its whole length. On
the dorsal edge, a pale shade broadest on the dorsal margin reaches two-thirds
of the way across to the basal streak. On the basal side of this spot, the ground
color seems a little deeper. Band III, although much angulated outwardly, is
not yet connected in the middle of the wing with Band

. . . i,7rjrr TV y VT ,,^^

IV; thus the first pair of white streaks are still con- ^^

nected and form an angulated fascia, broadest at its ^^

extremities. Bands IV, V, Viand VII remain entirely

separate and retain their primitive straight edges, with Fig. is. Later stage in the

the exception of the inner edge of Band IV which has "p^^^^* of color in the wing

^ ^ 01 L. ostrycefoLieila.

been somewhat produced toward the base at its ex-
tremities. Band III and Band VII are possibly a shade darker than the other
portions of the wing; the scales attached around the apex projecting into the
cilia are, however, paler and concolorous with Bands IV, V and VL

The ground color continues to darken, reaching its adult condition at a
relatively earher stage than is the case in L. cratcegella. At the same time {Fig.
15), the pale yellow spreads onto places originally unpigmented; thus the bands
become connected by a paler yellow shade, so that the white fasciae are trans-
formed into pairs of opposite narrow white streaks. At this stage in the develop-

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 131

ment, there are four dorsal as well as four costal streaks. The apical dot is
present, dark brown and almost as large as in the imago and very definitely
outlined. Around the extreme apex, the scales projecting into the ciha are
tipped with brown, this color being deepest just opposite the apical dot but
quickly fading out on either side, so that it does not reach to the last pair of
white streaks. The apical spot and these few scales around the apex are the
only scales on the wing which show the slightest sign of brown; the bands are,
however, a little deeper yehow on their outer than on their inner borders.

Very soon after the stage just described, the dark margin of the second dorsal
streak begins to develop; at this time, it consists of a few darker and shghtly gray-
ish tinted scales somewhat removed from the dorsal margin. The apical spot
is brownish black and seems to have reached its final development. The line
in the cilia, though darkest immediately opposite the apical spot, is plainly visible
from the fourth costal to the third dorsal streak, thus passing below the fourth
dorsal streak which at this stage has almost disappeared owing to the fusion of
Bands VI and VII.

Very little change is now needed to reach the adult condition. The margins
of the second pair of white streaks continue somewhat in advance of the others
and reach their final stage of development a little earlier. All trace of a fourth
dorsal streak finally disappears.

LithocoUetis lucidico Stella Clem.

At the time of the first appearance of a pale yellow tinge upon the wings,
it is possible to discern faintly four bands in the apical half of the wing. Just
before the apex within the area occupied by Band VII is a smah colorless spot.
With the exception of the fact that these four bands are not connected along the
middle of the wing, the adult color areas (Fig. 44, PL III) are now laid down.
The dorsal portion of Band IV, the most proximal of these four bands, is con-
tinued toward the base, as in the adult, by a pale yellow streak along the fold.
Its costal portion is preceded by a white streak bordered internally by pale
yellow which is extended along the upper median vein to the base.

The apical dot is the first dark mark to make its appearance. While the
yellow is still very pale, the scales of the formerly colorless spot near the apex
begin to take on a grayish tinge. The deepening of scales over this area seems
to take place in very much the same manner as that of the dark scales before the
apex in L. cratcegeUa.

The apical spot continues to deepen and increase in extent, reaching its
adult condition before there is any indication of the formation of dark margins.
The development of the apical spot is accompanied by a gradual and uniform
deepening of the ground color together with the fusion of all the bands in the
middle of the wing and of VI and VII on the termen. The middle portions of
the wing between the apices of the white spots are somewhat paler, showing
where the pigment has recently appeared.

132 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

The darkening of scales in the ciha around the apex and just below it on
the termen and the dark margins of the first costal and dorsal streaks are the
next characters to appear. A wing examined at this stage shows the scales in
the ciha brownish with scarcely a tinge of gra}^, the margin of the first dorsal
streak somewhat paler and brownish yellow, extending from the ciha to vein 2,
and the margin of the first costal streak pale and scarcely distinct. These
observations are in agreement with those made on L. cratcegella and ostrycefoUella;
namely, that the dark markings appear earliest in the apex of the wing on Band
VII and soon after on Band IV.

The process of development consists now merely of the gradual appearance
and darkening of the margins of the other streaks and the extension of the line
in the ciha to the first dorsal streak.

LithocoUetis morrisella Fitch.

At a period when the entire wing, viewed by transmitted hght, seems uni-
formly colorless, the white markings as seen by reflected light are definitely laid
down as they will appear in the adult (Fig. 48, PI. Ill) and have already acquired
the lustrous pearly tint characteristic of the adult. This white appearance is
due entirely to the structural modification of the scales described earher in this
paper and cannot be in any way ascribed to a precocious development of the
scales over these areas, since, when examined, all of the scales of the wing are
seen to be fully developed.

With the first appearance of pigmentation, a pale yellow suffuses those

portions of the wing only which are destined to be yellow in the adult {Fig. 16).

,,, ir V yt^yj, In the basal half of the wing, therefore, the yellow is

confined mainly to the region above the fold; there is,

however, a faint yellow tinge just before the dorsal arm

of the fascia. There is a scarcely discernible tinge of

Fig. 16. Early stage in gray at the basc of the dorsum below the basal streak

the development of color in ^^^ ^^^^^^ ^^iQ first dorsal streak. At the tip of the wing,

the wmg 01 L. morrisella. i o;

over the area to be occupied by the apical dot, the scales
are beginning to turn gray, those outermost being the deepest. Otherwise, the
areas occupied by gray and black in the adult and by the white streaks appear
by transmitted light uniform and colorless. Examination by reflected light
shows that these two kinds of areas are not homologous, since the white streaks
stand out clearly as structural modifications, while the scales margining them
and those below the fold are duller and lack the pearly luster. The bands and
the white fascia and streaks occupy their normal positions. The absence of a
white streak over the apex of vein 7 is accounted for when it is seen that the
apical spot has acquired such proportions as to cover the tip of vein 7, which in
this and related species reaches the margin nearer to the apex than usual.

In the next stage examined {Fig. 17), the ground color has deepened some-
what, and decidedly more of the grayish tint has developed. The dorsal part

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 133

of the wing below the fold, to be occupied by black scales in the adult, is pale
gray; beyond the fascia where the scales are merely brown tipped, the color is
still yellow, concolorous with the rest of the wing. Internal and external margins
of the streaks are beginning to form simultaneously, ^^^ ^^ ^

those of the first costal streak are the deepest. The /^^. -'- ^ ; w^w
change is direct from the colorless scales margining the .^ ^

white streaks to pale gray. The black streak in the fold

beyond the fascia is at this stage composed of pale gray Fig. 17. Later stage in the
scales. The apical dot is still pale gray, but the outer- development of color in the

^ ± CT .7 7 wmg oi L. momsella.

most row of scales is much darker and continuous on

either side with a line of gray scales extending from the fourth costal streak
to the third dorsal streak. The scales whose tips are to form the dark line in
the cilia are still white.

On the next wing examined, the gray has decidedly deepened along the dorsal
margin, but the golden color has not reached its final condition. The develop-
ment of the margins has progressed considerably, but now the internal are
decidedly in advance of the external margins. The former have a blackish color,
especially toward the costal margin; the latter are still grayish. The external
margin of the first costal streak is deeper than that of the others. The black
streak in the fold has practically reached its adult condition. The apical dot
has increased in size and color until its extent is now definitely defined. The
scales around its outer edge, which in the preceding stage were the only dark
ones, are now black. The line in the cilia is gray, but there is no indication of
the bluish iridescence.

From this time on, no new characters appear, the development consisting
only of a deepening of the colors to their adult condition.

The sequence of colors and the relative time of appearance of the various
adult characters is very much the same in L. robiniella as in the species just
studied. The apical spot and the black streak in the fold reach their final
condition at a somewhat earlier stage in comparison with the development of
the other markings.

Lithocolletis hageni F. and B.

This species (Fig. 17, PL III) possesses a well-defined apical dot, very decided
external margins to the white streaks as well as the usual internal ones and a
dark margin along the upper side of the basal streak. The margin along the
upper side of the white basah patch is regarded as homologous with the margin
of a basal streak, that along its outer side with the external margin of the white
streak primitively separating Bands II and III. The streak of ground color
beyond the basal white patch is, therefore, the dorsal portion of Band III.

Very few specimens of this species were available for study, but the obser-
vations confirm in general those made upon the five preceding species.

The apical dot appears well-defined and blackish at a period when Bands

lU

134 EVOLUTION OF COLOR PATTERN IN LITIIOCOLLETIS.

V, VI and VII are still separated from one another by curved white fascise
extending entirely across the wing and into the ciha {Fig. 18). The second
pair of white streaks are still connected^ but Band IV is beginning to project
i,jr,7n jjr ^- into them in the middle of the wing. At this time the

'^" ground color of the wing is a pale straw and lacks alto-
* ^ gether the golden brown of the adult. The internal
margins of the first and second dorsal streaks are begin-
FiG. 18. Wing of L. hagcni at ning to form. The inner edges of the bands^ border-
the first appearance of the dark -^^ ^^^ ^^^^^^^ outwardly, present much the Same

appearance as they do in species where the white
streaks never acquire external margins; that is, the ground color fades gradu-
ally into the white.

A dark line in the cilia begins to form soon after this; the scales showing
the brown pigment at this time are not those which will be bluish iridescent in
the imago; these latter remain pure white and only acquire their pigment later.

In the third specimen studied, the extent of the white streaks has been finally
limited, and the golden brown color of the adult has been attained. AU of the
internal margins of the streaks, the dark scales at the apex of the first dorsal
streak, those connecting the tips of the second pair of streaks and the two leaden-
colored lines of scales running across the wing from the tips of the third and
fourth costal streaks are present. In addition, external margins are present
on the white basal patch and on the first dorsal streak. All of these borders
are as they will appear in the adult. In this specimen, there was no indication
of a deepening along the upper side of the basal patch. However, this obser-
vation cannot be accepted as altogether reliable in indicating the time of appear-
ance of the margin of the basal streak, since the specimen was evidently ab-
normal, the apical dot not being present. No other specimens were available
to decide the point. Imagoes of this species are occasionally found which lack
both the apical dot and the dark margin along the upper side of the basal patch.

LithocoUetis aesculisella Cham.

The species heretofore studied have been those belonging to the group
having cylindrical larvae and characterized by the principal margins of the
white streaks being internal. This species (Fig. 70, PI. IV) and those following
have the so-called flat type of larva, and the white streaks of the imagoes have
their darkest margins on the outer side. That is, in the former group, the
tendency is for the color bands to acquire first, dark margins on their outer
edges; in the latter group, the condition is reversed, and the dark color appears
first on the inner edges of the bands.

In L. cBsculisella, at a period when there is scarcely any differentiation
between ground color and bands, the pattern appears to be laid down almost as
in the adult (Fig. 19). The bands are very pale buff, and judging from their
uniform color, appeared simultaneously. The pale shade at the base which

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 135

separates Bands I and II extends entirely across the wing, but the early fusion
of these bands on the costa is foreshadowed by the very faint tinge of buff just
making its appearance in the costal portion of the pale fascia. On the extreme
costa, Band V is indistinctly separated from VL Band VI is separated from
VII by a white streak extending entirely across the wing. With these exceptions,
the yellow marks are defined as in the adult. The white oblique dorsal streak
extends just within the termen, being separated from the white streak before the
apex as in the adult. The distal end of the oblique

dorsal streak is probably homologous with the middle ^ ,-<- ' '^zjj:^..:-'^^'^
portion of the white fascia which in primitive types ; . ""

separates Bands Vand VL In the imago, it sometimes

exists as a separate small white spot before the apical Fig. 19. Early stage in the
dusting and more proximal than the obhque costal development of color in the wing

^ ^ PI 01 L. oBSCuhsella.

streak. Along the termen, the cilia are of an unbroken

yellow color, showing that at the extreme margin, there has been a complete

fusion of Bands V, VI and VIL

Before any of the dark margins begin to appear, the ground color deepens
into the orange yellow of the adult. The external margin of the second fascia
and the dark dusting at the extreme apex are considerably in advance of the
other dark markings. The margin of the second fascia, formed by a dark line
of scales along the inner border of Band IV, is most distinct in the middle
portion of the wing and has not appeared on the extreme dorsal end nor on the
costal end beyond the angle. Its color at this stage is brownish. In one speci-
men, the apical scales are not as dark as the margin of the second fascia; in another
of the same age, as indicated by the state of development of the margins, the
dusting of the apex is somewhat darker than the margin of this fascia; in both,
however, the brownish tint is confined to the extreme apex. There is a slight
darkening of the scales in the middle of the wing on the inner border of Band III
adjacent to the first white fascia and along the under side of the oblique dorsal
streak.

The dark margins of the fascise and of the obhque dorsal streak continue
to develop, but before they reach their adult condition, the scales on the outer
side of the costal spot over vein 10 begin to darken. At this time, the apical
dusting is still confined to the apex of the wing and is not extended along the
termen toward the tornus. This extension of the dusting toward the tornus is a
characteristic of specimens appearing in the summer and does not occur in forms
from overwintering chrysahds. It develops after the other adult characters
have appeared.

Lithocolletis bethunella Cham.

In the youngest specimen examined, the ground color is a uniformly pale
ocherous, and upon it the extent of the adult markings (Fig. 84, PI. IV) is quite
definitely blocked out (Fig. 20). The white fascia (or second pair of streaks)

136 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

is, however, prolonged outwardly at its angle along the wing to a position between
the third pair of streaks. A comparison with the adult shows that the pro-
longation of the fascia is occupied by dusting on a whitish ground; that is, these
j.^^^ ni jy- y^y^ white scales are destined to become black tipped.

The third pair of streaks are situated normally over

I veins 10 and 2 respectively. There is no white spot

^ over the tip of vein 9 on the costa. Bands V and VI

Fig. 20. Early stage in the havc fuscd. A quadrate area in the apex remains

orZtZ*;°'"^ '-^ *'^ ^'"^ ««loriess; the portion directly over vein 7 corre-

sponds to the white streak normally situated here,
the remainder will be occupied by dark dusting in the adult. The scales
around the margin at the apex are yellow; these in the adult will be brown
tipped yellow scales.

In a wing at a somewhat later stage, there has been some deepening of the
ground color. There is a small patch of scales just beyond vein 7 which are
brown tipped and darker than any other scales on the wing. In the extreme
apex, the scales extending over the cilia are not noticeably dark tipped. Dorsal
to the dark patch of scales, the scales over the formerly colorless area become
gradually paler, until near vein 5 their tips show just a faint tinge of gray. All
the scales in the prolonged portion of the fascia are becoming pale gray tipped.
There is also a faint gray tinge to the scales just below the tip of the dorsal
arm of the fascia, continuous with the gray in the prolonged portion of the fascia.
All of these scales are much paler than those in the apex just beyond vein 7 but
approach in color those near vein 5.

The margins of the remaining streaks soon make their appearance; the
dusting of the apex reaches the adult condition earliest.

A contrast in the sequence of development of dark marks is afforded by
L. carycefoliella Clem. In this species, the margins of the first and second fasciae
develop contemporaneously with a small patch of dusting at the extreme apex,
and the margin of the second fascia is entirely complete, though not as dark as
in the adult, at a stage when there is no indication of dark dusting beyond its
angle. This dusting is very variable even in the adult, sometimes being very
slight. The significance of the difference in development will be discussed later.

LithocoUetis ulmella Cham.

In this species, the pale yellow color appears simultaneously over all those
areas which are to remain yellow in the adult (Fig. 95, PI. IV) or in which the
scales are yellow with brown or blackish tips. The white streaks and the white
dusted portions of the wing remain colorless. Thus, the second costal streak is
produced for a short distance along the cell, and the wing from vein 7 to the
apex is colorless, but the scales around the apex and below the oblique dorsal
streak, which are yellow with brown tips in the adult, have already acquired the
pale yellow color.

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 137

In the next stage studied^ the color has become almost as deep as in the
adult. A number of the scales projecting from the extreme tip of the wing over
the cilia are brown tipped; these are continued in a broken line through the cilia
to the tornus. Over the wing membrane near the apex, in the area which was
transparent in the earlier specimen, the scales are becoming dark tipped. These
scales are deepest toward the apex of the wing, becoming very pale gray toward
the tornus external to the oblique dorsal streak. A few scales along the costa
corresponding in position with the dark scales in the adult are beginning to show
darker tips. The scales margining the first and second costal streaks, which were
whitish before, are now faintly gray. The color is uniformly developed over the
whole margin and on the prolonged portion of the second streak.

In a specimen a little further advanced than the one just described, the apical
dusting is almost as in the adult. This darker color does not quite reach the
tornus; that is, the external edge of the white dorsal streak is stih margined by
pale gray scales as before. The margins of the first and second costal streaks
have scarcely changed in color, but the scales in the prolonged portion of the
second streak have darkened more rapidly; they are decidedly gray but much
paler than those in the apex. The only new character which has appeared is
the dark line through the cilia.

Lithocolletis hamadryadella Clem.

In the earliest stage examined {Fig. 21)^ the basal third of the wing is almost
colorless. At the extreme base on the costal margin is a small patch of pale
yellow scales, and a little beyond it another similar patch. These two patches
of scales form Band I. Beyond these is a somewhat larger patch of pale yellow
scales. This is the inner portion of Band II; the line of scales which marks its
original outer limit in the adult (Fig. 77, PL IV) has not yet acquired any pig-
ment. Band III at this stage is represented by a rather narrow angulated band
which extends across the wing some distance before the tip of vein 16. The
white projects into its angle on the inner side. Band IV is composed of two
separate patches of yellow scales, one on the costa and the other on the dorsum.
Their pointed apices lie over the bases of veins 10 and ^^^

2 respectively, the normal position for Band IV, but - "^^ ^ ^^^

on the margins, they curve back toward the base so >

that they are much farther removed from the tips of
the veins than usual. The middle of the wing to the ^^

apex is colorless. Band V + VI is represented by a ^^^- ^i- Early stage in the

, 1 r* 1 'ii • rrn 1 1x1 development of color in the wing

patch of scales on either margm. The dorsal patch of l. taadr^a^a.
originates just beyond vein 2 and extends over vein

5; the costal patch originates on the cell about midway between veins 10 and
9 but extends obliquely outward so that the inner edge on the costa is just basal
to vein 9 and the outer edge just before vein 7. The white streak over the tip of
vein 7 is very narrow, and its corresponding dorsal streak is indistinct. The

138 EVOLUTION OF COLOR PATTERN IN LITLIOCOLLETIS.

cilia belonging to Band VII are all yellow with a few white scales here and
there extending into them. Bands VII, V + VI and IV are somewhat deeper
than Band III; Bands I and II are very pale, their paleness being caused by
the large admixture of colorless scales at this stage.

In the next specimen examined, apart from a slight deepening of Band III
and those beyond it, the only changes observed are the following. The two basal
patches on the costa, belonging to Band I, show a slight admixture of brownish
tipped white scales. The third patch (Band II) is now faintly continued to

the dorsum by a line of scattered yellow scales, and just

1 ^^ ^'^^^ m beyond it, in the position normally occupied by the

'7 '^'"'^ outer margin of Band II, is an angulated line of very

faintly yellow scales. This marks the original outer

, -' limit (phylogenetically) of Band 11. Just before the

Fig. 22. Later stage in the tip of vciu 16 is a similar line of scalcs running parallel

development of color in the wine , ,i i i r* t~» i ttt i i • -j i

of L. hainadryadeiia. ^^ ^^^ dorsal arm 01 Band III and markmg its outer

limit. These two lines of scales sometimes do not ap-
pear until a much later period in the development. Their time of appearance
is largely determined by the number of yellow scales mingled with the dark
scales in the adult, since it is only in the yellow scales that the pigment ap-
pears early.

In the next stage studied (Fig. 22)^ Bands I and II in the basal part of the
wing remain practically unchanged. Band III is the same in extent, but its
color has deepened. Its internal margin is beginning to form, the scales adjacent
to the yellow becoming gray tipped. The gray pigment appears directly in
the whitish uncolored scales of the previous stage without passing through any
of the intermediate steps. In general, the margin is broadest on the costa, be-
coming narrower and paler below it, and not easily distinguishable toward the
dorsum. On its inner edge, the margin is straight or slightl}^ curved as in the
imago; on its outer side, the scales project into the angle of the band and fill
up the formerly colorless space. The tips of these scales are darker than those
of the margin elsewhere. In this specimen, there are only a few scattered yellow
scales beyond Band III and no indication whatever of the dark line of scales
which is to appear later. The internal margin of Band IV is well marked on its
dorsal half, consisting of several rows of dark tipped scales as in the imago. It
curves between the two yellow patches, extending to their apices, but not beyond
them into the colorless fascia. The costal half of the margin consists only of a
few pale gray scales and does not quite reach the dorsal half of the margin.
Except for a few very faintly grayish scales above the termen on the inner edge
of the dorsal part of Band V + VI, there is no indication whatever of any dark-
ening of scales adjacent to Band V + VI. Belonging to Band VII, and begin-
ning just behind vein 7, the scales covering the entire apex of the wing are gray
tipped. Some of them, especially on the costal margin, project over the yellow
scales which are to form the line in the middle of the cilia. The gray tipped

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 139

scales of the apex, those between the two parts of Band IV and in the angle of
Band III. are darker than the other scales forming the margins.

In a specimen at a considerably more advanced stage of development, the
scales composing the two patches at the base of the costa are entirely gray tipped
and very much the same as they will appear in the adult. The colorless scales
mingled with those forming the proximal part of Band II are beginning to be
gray tipped; this is especially noticeable on the quadrate part above the upper
median vein; below it, there is only a line of a few scattered gray tipped scales
reaching to the dorsum. The line of scales beyond, forming the original outer
limit to Band II, is pale gray tipped throughout and, though broader on the costa,
is complete to the dorsum. The broadening of the internal margin of Band III
has continued. Its external margin, separated from the yellow band by white,
is indicated by a single line of pale gray tipped scales running from the costa to
the cell. The internal margin of Band IV has broadened and deepened and has
now probably reached its adult condition. It is not possible to determine
whether the development is completed in any single specimen, since there is a
great deal of variation in the amount of dusting among the different individuals
in this species. In this stage, the scales which in the imago are the most distal
of those forming the dark external margin of Band IV are tipped with pale gray
and connect with the dusting between the costal and dorsal parts of Band IV.
Those on the inner side of the margin which will in a great measure fill up the
white space between the margin and the yellow parts of the band are still un-
colored. There is a very considerable patch of dusting internal to Band V + VI
on the dorsal margin and connected with the apical dusting which has been

IV

V+VI

extended proximally between the parts of V + VI. ri l^~ „ ,.,,

The costal dusting is not nearly as dark and is some- ^-"^^^^"^ ^f^^^.

what scattered. In addition to the extension of the

apical dusting proximally, there has been an increase

in the number of dark scales mingled with the cilia. Fig. 23. Wing of l. immadry-

The dark line through the middle of the cilia formed ^^'"^ ^^^^ ^^^ ^^^ ^^ ^^^ p^p^^

state

by brown tipped yellow scales is now quite distinct.

At this time, most of the dark markings are at least definitely defined in posi-
tion; there is needed to bring them to their final state only a darkening of those
scales already gray tipped and in some places the appearance of pigment in
other scales adjacent to these.

In a specimen to be described now, most of these characters have appeared
{Fig. 23). The patches of dark gray scales at the base of the costa and the hne
of scales beyond the third of these patches have increased and darkened decidedly.
The line of scales external to Band III has now become dark gray tipped. Ex-
ternal to Band IV, the dusting has been extended toward the base, so that only
a small white streak is left between it and the band on the margins. The dusting
in the apex has been much increased and extends along the margin, especially on
the costa, and connects with the internal dusting of Band V, thus almost obhter-
ating the white streak over vein 7.

140 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

Cremastobombycia ignota F. and B.

In the youngest specimen examined^ one in which the ground color was pale

yellow^ the extent and position of all the adult markings (Fig. 62, PL IV) is defined

{Fig. 2Jj) . The paleness is especially marked toward the apex, where it is difficult

r,r to distinguish the markings at all. The white fascia

III l^ V+J/I [77 .

^^^^ is considerably broader than it will appear later. A

careful examination shows it to be composed of two
parts, an inner fascia which is to form the white fas-
cia of the adult and in which the scales are whiter

Fig. 24. Early stage in the de- ^j^^ ^^^ f^^y formed, and an outcr fascia, where the

velopment of color in the wing of , ,1 n i i i i

(7 ignota. scales are not as large nor as well developed and are

somewhat buffish in color. These scales will later
form the black scales of the margin. This same thing is true of the other
streaks but is not as noticeable except in the case of the first costal because of
their more irregular shape. The band between the pair of streaks at three-
fourths of the wing length and the streak enclosing the apex is, as shown by its
position, equivalent to Band V + VI. The white streak before the apex is situ-
ated over the tip of vein 7.

In a later specimen, the ground color has deepened to that of the adult.
The scales in the apex and in the cilia remain concolorous with the wing and are
not darker at their tips. The dark external margins of the white fascia and
streaks are appearing simultaneously, and all are equally advanced. The
scales forming them correspond in position with the colorless scales which formed
the outer part of the fascia in the preceding stage. The pigment, which is pale
gray without a tinge of yellow or brown, is evenly distributed throughout the
entire scale from the base to the tip. None of the internal margins has appeared.

In a later stage, the external margins are black as in the adult, and internal
margins are present on the costal end of the first white streak and the fascia.
There are no internal margins on the pair of streaks at three-fourths. As
margins of these streaks are not always present in the imago, it is impossible to
determine conclusively whether they normally appear after the internal margins
of the fascia and the first streak. However, as will be shown later in the dis-
cussion, variable characters tend to appear later in the ontogeny than the fixed
characters of a species.

The observations on the development of color in the wings of C. solidaginis
(Fig. 60, PI. IV) agree very closely with those on the preceding species. The
configuration of the color areas at their first appearance is the same as that of
the adult. In this species, however, the scales which are to form the dark
margins in the adult are pale yellow instead of, as in ignota, being colorless
and not fully formed. A comparison with the adult shows that these are
dark tipped yellow scales, the dark pigment being found at the tip of the scales
and not, as in ignota, almost uniformly distributed throughout the entire
scale.

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 141

4. Discussion and Conclusions.

In general, I have found that the action of three processes is sufficient to
explain the transformation of the seven primitive bands into the several existing
patterns of the ground color. These three processes may be stated thus: (a) The
middle portion of a band may be produced distally until it comes in contact with
the band beyond it. (6) The extremities of a band may be broadened by being
produced proximally. (c) The extremities of a band may be narrowed by the
retraction of pigment from their outer edges. These three processes may act
separately or in combination; the particular pattern produced is the resultant
of their combined activity. The first two of these laws are based chiefly upon
the changes observed during pupal development, the third law is deduced
from adult characters, since the changes are such as cannot be observed directly.
The evidence in detail is as follows :

In L. tiliacella {Fig. 7), seven distinct transverse bands of ground color
are laid down; of these, those numbered II, III and IV appear first; following
these, Band I, at the base of the wing, appears continuous in the costal half of
the wing with II; it is, however, sufficiently separated from it by the difference
in time of appearance. The three bands in the apex of the wing appear last and
undergo no change; the connection of Band V with IV is brought about entirely
by the extension outwardly of IV along the middle of the wing. The changes
in Bands II and III are brought about by the spreading of the color onto areas
formerly unpigmented; that is, the fascia between II and III is narrowed by
the very slight growth distally of II along the middle of the wing, its extremities
remaining fixed, and by the extension proximally of III, chiefly along the margins.
Band IV reaches its adult configuration earhest; it is also the earhest to acquire
the black pigment on its outer edge. This does not appear in the middle of the
wing, where the band has extended outwardly uniting with V. The outer
edges of the extremities of Band III have remained constant in position, and the
dark scales appear first here. In Band II, there has been the least change in
shape; the deepening of the pigment along its outer edge occurs more uniformly.
From this, it follows that while the general tendency is for dark pigment to develop
along the outer edges of hands adjacent to unpigmented areas, the precise time of its
appearance along any hand or portion of a hand in relation to other hands or portions
of the same hand is determined hy the time when such hands or portions of hands
reach their adult shape and color. The outer edge of a hand in direct contact with
another hand acquires no dark pigment. It would appear from the observations
on this species that the hands in their primitive condition are narrower than they
are in the adult. A fascia may he narrowed hy the uniform spreading of a hand
toward the hase.

These conclusions are corroborated by the observations on L. tritcenianella
(Fig. 5, PI. Ill) ; all the bands, however, appear simultaneously, and at this time
all are straight. There is, however, a reduction in the apparent number of bands,
V and VI being completely united. Band IV seems no further advanced than

142 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

the others toward the adult condition; as in tiliacella, development must take
place more rapidly, since its margins reach their adult condition before those
of the other bands. Here, also, is the first manifestation of that tendency, so
characteristic of all the more highly differentiated species of the genus, toward
the darkening of scales in the apex. In this species, these scales are scattered
over the entire apex, instead of being limited to a definitely defined area where
the scales are modified structurally. This species does not, at any stage, show
the primitive type of marking; a considerable amount of development is, however,
necessary before the color areas attain their definitive final extent.

In L. cratcegella (Fig. 36, PI. Ill), ostrycefoliella (Fig. 23, PL III), lucidicostella
(Fig. 44, PI. Ill), hageni (Fig. 17, PL III) and morrisella (Fig. 48, PL III), in the
basal half of the wing, where in the adult there is the greatest divergence from a
transversely banded type of marking, the color areas, upon the first appearance
of a uniform pale yellowish, have almost exactly the same shape and extent as
in the adult. They undergo very little change during development, except that
in these species, as weh as in tiliacella and tritcenianella, there is a progressive
and gradual deepening of the ground color. In the youngest specimens of
ostrycefoliella observed (Fig. I4), the first pair of streaks are still connected;
very soon, however, they become separated by the union of Band III with IV.

In the apical half of the wing. Bands V, VI and VII appear as straight
transverse bands (except in morrisella) ; the transformation of the unpigmented
fasciae between them and between IV and V into pairs of opposite streaks is
brought about by the outward extension of Bands IV, V and VI along the cell
in a manner similar to that previously observed for Band IV in tiliacella and
tritcenianella. The obliteration of a white streak is brought about by the inward
extension of the extremity of a band to meet the outer edge of the preceding
band. In morrisella, the pattern is predetermined from the outset, the white
fascia and streaks being produced by structurally modified scales, which are
destined never to acquire pigment.

There are no reasons favoring the view that such a type of marking as that
found in cratcegella (Fig. 36, PL III) is primitive: there are many reasons for re-
garding it as having been produced through a long process of evolution. It is a
matter of general observation that the pale colors appear first ontogenetically,
black of the pigmental colors being last to appear; it has been inferred that the
same is true also for the phylogeny. In the adults of these species, there have
developed, in addition to the usual dark markings adjacent to the unpigmented
areas, definitely limited black spots and streaks in the apex. The scales thus
pigmented are also differentiated structuraUy from the surrounding scales by
their smaller size. There is no such structural modification of the dark tipped
scales in the apex of tritcenianella. Obviously a considerable period of time
must have been required for the final differentiation of such characters as these.
That these species have undergone an extensive process of differentiation in the
production of certain characters of a kind which can only be acquired late in

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 143

racial development, renders it highly probable that there has also been a far-
reaching modification from the ancestral condition of the extent of the color
areas. These phylogenetic changes are scarcely or not at ah repeated in the
ontogeny. Three primitive bands are still preserved during the early stages
in the apex of the wing; the band proximal to these shows unmistakable signs
of its origin from a straight transverse band, its outer edge being straight. We
may, therefore, with reasonable certainty, conclude that the color pattern of
the basal part of the wing may be traced back through a series of phyletic changes
to its origin from the first three primitive bands.

The ontogeny offers no direct testimony of the means by which these results
have been accomplished; it has given, however, visible evidence of two actual
dynamic methods by which changes have been wrought; namely, the middle
portion of each hand may he produced distally until it comes in contact with the
hand heyond it; the extremities of a hand may he hroadened hy heing produced proxi-
mally. Thus are produced, first, angulated fascise; second, slightly obhque
pairs of opposite streaks; the final effect of the second of the above processes, if
acting alone, is the entire obhteration of the white streaks. That a third process
has been at work is indicated by observations upon adult forms. The dorsal of
the first pair of white streaks in L. cratcegella (Fig. 36, PI. Ill) does not occupy its
primitive position over vein 16 but has been displaced toward the base, and vein
16 reaches the margin within Band IV; that is, there has been an extension of
the extremity of Band IV in this direction over the area formerly white. But
since a white streak still remains, there must have been a compensating drawing
away of the outer edge of Band III at its extremity. Such a withdrawal of
pigment cannot, of course, be observed ontogeneticahy. A third dynamic process
in the transformation may, therefore, be stated thus : the extremities of a hand may
he narrowed hy the retraction of pigment from their outer edges. This is an example
of the shrinking away of a band at one end. In this case, the shrinking takes
place at the outer edge of the band only. The result of the combined action of
the second and third of these processes is the progressive movement of the
end of the band toward the base.

These three processes are sufficient to explain the origin of all the types of
marking in this genus; the specific results in any case are due to the manner and
extent of their action; that is, whether they act singly or together and to the relative
rapidity of action of the three processes. How they have combined to produce
these various types will be discussed in detail under the heading of phylogenetic
development.

In the second division of the genus, the extent of the areas of ground color
undergoes very little change during development. The ground color appears
in the earhest stages over those areas which in the adult are occupied by yellow
or reddish yellow scales or by dark tipped yellow scales. The scales, which in
the adult are white with dark brown, gray or black tips, do not receive pigment
until much later; they then pass directly to gray or brownish gray and thence to

144 EVOLUTION OF COLOR PATTERN TN LITHOCOLLETIS.

black, without passing through the intermediate yellowish and brownish shades.
The tendency for dark pigment to develop along the edges of bands adjacent to
unpigmented areas is exemplified in this group also, with this difference, however,
that it appears first along the inner edge of a band. When dark pigment develops
in addition along the outer edge, it appears at a later period. The converse is
true in the first division of the genus; the dark margin appears earlier on the
outer side of a band. I can offer no explanation of the original cause of the
difference in direction of development; the result in either case seems to be the
expression of the same general tendency in development, acting, however, with
different degrees of intensity on either side of a band. It would seem that the
development of unusually dark pigment on one side of a band — the acceleration
of chemical processes over a localized area — retards development on the other
side.

The ontogeny within this division affords very little evidence of the operation
of the first two processes of evolution defined earlier in this discussion. However,
the shape of the bands and the separation of the white fasciae into opposite spots,
with the total obliteration of some of the spots, is sufficient proof that these forces
have acted during the evolution of the species.

L. hamadryadella (Fig. 77, PL IV) alone, of those observed, shows the out-
ward prolongation of the middle of a band; the gradual extension distally of
the dark dusting between the dorsal and costal portions of Band V + VI and
its final contact with the apical dusting (Figs. 22, 23) is an example of the
working of this law. The third law postulated to explain the displacement of
oblique streaks toward the base receives decisive confirmation from the adult
markings of the same species. A narrow line of scales crosses the wing beyond
the yellow portions of each of Bands II, III and IV, being separated from them
by white. These scales occupy positions corresponding to the outer margins of
these bands; ontogenetically, the dark pigment develops in them after it has
appeared on the inner margins of these bands. The conclusion is apparent that
these lines of scales mark the earlier phylogenetic limits of these bands. The
bands have shrunk away, most rapidly on the costa and dorsum, leaving behind
them the margins as visible evidences of the process.

In L. tiliacella {Fig. 7), which, from its conservation of the greatest
number of bands in a primitive straight condition, the very pale ground color
and the entire absence of any dark scales in some of the bands, may be regarded
as the most primitive of the American species, there is a difference in the time
of appearance of the bands. In the more advanced species, the bands appear
at the same time. Evidently, then, in the higher species, there is a hastening
and crowding together of the earher stages, so that the bands which formerly
appeared in sequence now appear contemporaneously. Those in the apex of
the wing, which have experienced the least modification during the phylogenetic
development of the species, alone recapitulate the racial history during their
ontogenetic development.

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 145

The discussion heretofore has dealt chiefly with the processes which effect
the final limitation and configuration of the ground color. The general tendency
toward the production of dark pigment in rows of scales adjacent to unpigmented
areas has also been noted. These streaks of dark scales, which are usually
transverse, and the various other characteristics, such as the apical spot, the
longitudinal streak of black scales in the apex and the black streak in the fold,
properly constitute the markings^ in distinction to the ground color, and are
superimposed upon it. While the progressive change in the configuration of
the ground color is the important factor in determining the general evolution of
the genus and the different stages at which development has halted form the
starting points for the differentiation of the different groups of species, the develop-
ment of the markings is the important factor in determining the phylogenetic
sequence of the individual species within each group. A knowledge of their
mode of origin and relative time of appearance is therefore essential.

In L. tiliacella and tritcenianella, it was seen that the dark margin develops
first along the edge of Band IV, which is the first band to reach its adult shape.
In L. tiliacella, the outer edge of this band has become fixed before all of the bands
have even been laid down. The dark margin appears toward the costa and
dorsum, in places where the outer edges of the band have undergone no change
whatever since the first appearance of color. The dorsal margin seems to be
in advance of the costal. There is no darkening of pigment along the bands in
the apex, which appear comparatively late. From these observations, it appears
that in the ontogeny, the dark margin develops earliest along the edge of that band
which first reaches its definitive adult condition. Before dark margins can develop,
the edge of a band must have remained fixed for an appreciable period of time.

In L. cratcegella, ostrycefoliella and lucidicostella, the dark margin on Band
IV is also the first to appear. The apparent contradiction to the principle just
laid down that a dark margin develops earhest along the edge of the band that
has first reached its definitive adult condition, lies in the fact that in these species
the color areas in the basal half of the wing are at the outset defined almost as
in the adult, while the bands in the apex repeat the racial development. It
might be expected, then, that the margin would appear earhest along the first
pair of streaks. It has, however, required a much longer time in the evolution
of the species for the basal half of the wing than for the apical half of the wing
to acquire its present aspect; the development in the apical half has come to a
standstill comparatively early in the phylogenetic history of the species. The
observed recapitulation in the apical three or four bands during the ontogenetic
development of the steps in the racial differentiation of the species is in harmony
with the conclusions of other investigators that it is only among primitive forms
that anything approaching a complete repetition of phylogenetic changes is to
be found in the ontogeny.

The apex of the wing (in respect to the conservation of primitive color
areas) may be regarded as primitive in comparison to the base of the wing. In

146 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

the phylogenetic development of this type of marking^ therefore, it is possible
that the margins appeared on Band IV while the basal half of the wing was still
under the influence of dynamic forces producing constant changes in the out-
Unes of the areas of ground color. No margins could have appeared in the basal
half of the wing until after these outlines had become permanent. Some of
these changes, as has been already shown, are such as cannot be repeated during
pupal development. Hence the color areas in the basal half of the wing tend
to be laid down in their final shape. The sequence of the appearance of dark mark-
ings, which are in reality a second series of transverse hands superimposed upon an
earlier set, follows the same order for the ontogeny as for the phytogeny; that is, in
this second and more recent set of characters, there is an actual recapitulation. The
margins of Bands V and VI appear later, corresponding, therefore, to the sequence
in the time of appearance of these bands phylogenetically, hypothetically based
upon that of L. tiliacella. The margin on the upper side of the basal streak
corresponds in time to the outer margins of the bands; it appears before inner
margins develop. An outer margin of a band tends to appear first near the costa
or dorsum, an inner margin near the middle of the wing; that is, at the apex of
the outer side of a white streak, these being the parts of the edges of the bands
which have reached their permanent position earliest. Therefore, the relative
time of appearance of the dark margin of any band in the ontogeny is dependent
upon the time when the edge of that hand hecame fixed in the phytogeny.

InL. morrisella (Fig. 48, PL III), margins are formed almost simultaneously
on either side of the white streaks by a direct change to gray and thence to black.
The pigment in the other black scales develops in a similar manner without
passing through the usual intermediate stages. The gray and black pigments
make their first appearance at an early stage, while the ground color is still
pale yellow. The structural differentiation of the white scales, the massing of
black scales in the dorsal half of the wing, the exceedingly large apical spot and
the black longitudinal streak in the fold all indicate the very high phylogenetic
position of this species. Upon the addition of these newer characters, the earlier
ones were crowded closer together and pushed farther back into the ontogeny,
resulting in their contemporaneous development.

In the species of the second division of the genus, dark margins develop in
conformity with the principles already enunciated. There are certain minor
differences between this and the preceding division attendant upon the original
unexplained diversity in the inherent tendencies of the two divisions. The dark
margin appears first on Band IV in the middle of its inner edge, in contrast to its
appearance in the other division on the outer side and near each end of this band;
the same is true for Band III. This is a necessary consequence of the action
of the three general processes of evolution; the middle of the inner edge of any
hand remains stationary, and hence a margin will develop here earlier than at the
extremities of the hand.

In Cremastohomhycia ignota (Fig. 62, PI. IV), the development is very similar

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 147

to that of the species of the '^flat-larval group.'' The margins appear almost
simultaneously^ the scales changing directly from colorless to gray and thence to
black. These scales, at the time the yellow ground color spreads over the wing,
are not fully formed, thus contrasting with the similarly unpigmented, but
fully developed scales belonging to the white fascia and streaks. In all other
species examined, the scales were all equally and fully formed at the time of the
first appearance of pigment in any part of the wing.

In many species, these dark streaks adjacent to white are the only defined
dark markings present; there may be a few scattered, darker tipped scales in
the apex. An increase in the amount of dark pigment producing additional
dark markings manifests itself in either of two ways. (1) The number of scales
thus affected may be increased, so that the actual extent of the wing occupied
by dark tipped scales is greater. This is the direction that evolution has taken
in the species of the '' flat-larval group'' (illustrated on Plate IV) and in a few of
the species of the '^ cylindrical-larval group" {e. g., celtisella, Fig. 56, PL III, and
celtifoliella, Fig. 4, PL III). Often, among the more highly differentiated of the
species of the '^ flat-larval group," the entire apex of the wing is covered with
black tipped scales {e. g., agrifoliella^ Fig. 79, PL IV). Similar areas may develop
in other parts of the wing; thus a patch of dark tipped scales may extend outwards
from the angle of a fascia (e. g., bethunella, Fig. 84, PL IV). (2) There maybe a
concentration of pigment in comparatively few scales, which with increasing
differentiation tend to be segregated in definitely limited areas marked out by
structural modifications of the scales (e. g., lucidicostella, Fig. 44, PL III, and
cratcegella, Fig. 36, PL III). In these cases, the scales are pigmented almost to
their bases, not merely dark tipped. With the single exception of obstrictella
(Fig. 64, PL IV) among the species of the second division of the genus, this
tendency is confined to the species of the '^ cylindrical-larval group," where it
manifests itself in the apex either as an apical dot or as a somewhat elongate or
irregular patch of scales, broadest just before the tip of the wing. A similar
longitudinal streak of black scales is present in the fold in robiniella (Fig. 50,
PL III), uhlerella (Fig. 49, PL III), and morrisella (Fig. 48, PL III).

Finally, there may be a decided deepening of the ground color itself between
the white streaks. This process has reached its highest development in robini-
ella, uhlerella, and morrisella.

In the ontogeny, the development of all of these specializations has been
much abridged, and concomitant with this, their time of appearance has been
pushed back farther and farther into the earlier stages of pupal development.
The dark tipped scales remain colorless during the time the yeflow scales are
attaining their adult condition; later, they change directly to gray and black.
The pigment in the areas of dark tipped scales in the apex and beyond the fascise
often appears simultaneously with that of the margins. Some variation is to
be noticed here; in L. bethunella, the gray pigment appears in all the scales
beyond the fascia, simultaneously with the beginning of the formation of the dark

148 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

margin; in L. carywfoliella, the dark margin of the fascia is complete before any
pigment has developed in the area beyond it. In the imago, the patch of dusting
beyond the fascia is an estabhshed constant character in hethunella; in carycejo-
liella, the amount of such dusting is very variable, indicating its recent appearance
in the species.

It would seem then that only those characters permanently established and of
long standing in a species exhibit this precocity of development; recently acquired or
variable characters appear ontogenetically in the order of their phylogenetic sequence.

This conclusion is confirmed by observations upon the mode of development
ontogenetically of other analogous dark markings. Comparison of imagoes
indicates that the apical spot and the black patch in the apex have originated
phylogenetically earher than the streak of brown scales along the middle of
the wing but later than the dark internal margins of the white fascise and streaks.
In L. crata^gella, the pigment in the scales of the black patch before the apex,
which is unquestionably a characteristic only acquired through differentiation
extending over a long period of time, appears earlier than that in the brown
scales extending along the middle of the wing and reaches its definitive adult
condition while the latter are still wholly unpigmented. In L. hageni also, the
apical spot appears sooner than the dark scales in the middle of the wing beyond
the tip of the first dorsal streak.

Such specializations as these cannot, however, have originated very early
in the evolution of a species, and the observed phenomenon of their unduly
precocious appearance and rapid development in the ontogeny requires additional
explanation. In L. cratcegella, ostryoefoliella, lucidicostella, hageni and morrisella,
these were the first dark markings to appear. The first tinge of gray appeared
before the ground color itself had reached its adult shade or rarely even before
it had reached its adult configuration. It is possible that physiological causes
act directly toward this end. In the racial development, the segregation of
black scales over a limited area and the later modification of these scales structur-
ally has been accomplished gradually; the results have become permanent in
the species, and the scales, when first formed before any pigment has entered
them, have a structure identical with that acquired late in phylogenetic develop-
ment. It is suggested, in explanation of the phenomenon of the early appearance
of black pigment in them, that the concentration of pigment-forming substances
within a smaller space — the interior of these smaller scales — results in a more
intense chemical action over a locahzed area and thus hastens development.

The general conclusions which have been drawn from the study of pupal
development may be summarized as follows :

1. The primitive type of color pattern consists of a series of seven trans-
verse bands, which, except the first, cross the wing perpendicular to the costal
margin and are not wider than the unpigmented fasciae between them. Their
positions are determined by the course of the longitudinal nervures.

2. In the evolution of the genus, changes in extent alone of bands are

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 149

brought about by the uniform spreading of the pigment toward the base, the
edges of the bands remaining straight. Changes in the extent and form of these
bands have taken place through the action of any or all of the three following
dynamic processes : (a) The middle portion of a band may be produced distally
until it comes in contact with the band beyond it. (b) The extremities of a
band may be broadened by being produced proximally. (c) The extremities
of a band may be narrowed by the retraction of pigment from their outer edges.
The action of the first two may be observed directly, the third is an inference
from a comparison of adult markings.

3. The bands, either in their primitive shape or as modified through the
course of evolution, constitute the ground color upon which a second darker
series of transverse characters is superimposed. These are properly termed the
markings and appear at the limits between ground color and unpigmented areas.
Their relative time of appearance, ontogenetically, is dependent upon the time
when that portion of the edge of a band has become fixed phylogenetically.

4. Later in the ph3dogenetic history, spots and longitudinal black markings
may appear. For reasons not fully understood, certain of these markings appear
earliest in the ontogeny.

5. Recapitulation in the ontogeny is at most only partial and is confined
to those species or to those parts of the wings in which there has been the least
advance from a primitive condition. Here also a distinction must be made
between ground color and markings.

(e) Phylogenetic Development of the Color Pattern.

From the previous discussion, it was seen that a few definite processes,
identical in all of the species, have acted to produce modifications in the extent
and configuration of the areas of ground color. Later, a set of darker transverse
markings has been superimposed upon the ground color. The dark pigment
appears in rows of scales contiguous to unpigmented areas. There are two
possibilities in regard to the place of its first appearance, namely, the outer or
the inner edge of a band.

While the evolution of the pattern of the ground color has taken place in
the same directions in both divisions and in the two subgenera of the genus,
the development of the dark markings has followed different directions. In the
division whose larvse are of the ordinary cylindrical type and in Porphyrosela,
the dark markings appeared first on the outer edges of the bands: in the division
whose larv2e are of the modified flat type and in Cremastobombycia, the margins
appeared first on the inner edges of the bands.

Two different inherent tendencies in development are therefore exhibited.
These two lines of development must have diverged early in the phylogeny,
while the only color pattern was still that blocked out on the white wing by the
shape of the areas of ground color and before any of the dark markings had
appeared, since neither type of marking can be derived from the other, and the

150 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

species of the first division are more closely related among themselves than to
any of the species of the second division or of Cremastobomhycia.

An ancestral form, more primitive in respect to generic structure as well as
color pattern and transversely marked with seven bands of a uniform pale
yellowish color, has evidently given rise first to a series of species in which dark
margins have developed earliest on the outer edges of bands — in this case, the
white streaks and fasciae are said to be internally dark margined. Later, or
perhaps contemporaneously, in another species of the same ancestral strain,
the opposite tendency has developed, and there has been evolved a series of
species in which dark margins appear earliest on the inner edges of the bands —
in this case, the white streaks and fasciae are said to be externally dark margined.
A few of these latter species, constituting the subgenus Cremastobombycia, have
retained a more primitive generic structure; the remainder have followed the
same course of evolution structurally as the species of the first division of the
genus. It is, of course, possible that this second tendency originated twice
and independently.

Several factors combine to render it probable that Cremastobombycia (in
its modern form) and the ''flat-larval group'' originated at a somewhat later
period than the typical LithocoUetis (Fig. 26). The restricted geographical
range of the ''flat-larval group'' and of Cremastobombycia, which with one
exception are confined to America, indicates their comparatively recent origin.
Considerations based upon larval characters confirm this view. The later larval
stages of the typical LithocoUetis are identical with those of Gracilaria, which is
without question accepted as the ancestor of all these groups. The corresponding
stages in the larva of Cremastobombycia have a more flattened and triangular-
shaped head, without, however, the structural modifications which have de-
veloped in the "flat-larval group." It may be concluded, then, with reasonable
certainty, that after the origin of the typical LithocoUetis, the species of the
ancestral stock retained for a short period their imaginal structure. During
this time, changes in the larval form were initiated. In the ancestral form of
Cremastobombycia, development along this line did not progress far, and the
larvse have preserved the normal mode of feeding; hence there has been no
attendant structural modification.

Cremastobombycia may have originated after a certain amount of evolution
in the color pattern had already taken place; this view is supported by the
observation that none of its species conserves as primitive a type of marking as
is to be found in several species of the "flat-larval group."

We may conclude, therefore, that the immediate ancestor of all the groups
under discussion was of a somewhat more primitive structure than Cremasto-
bombycia as we now know it. The structural relations of the four groups with
reference to the ancestral stem are illustrated diagrammatically by Fig. 25; the
relative time of origin of each by Fig. 26. The similarity of the color pattern of
the species of Cremastobombycia to that of some of the species of the second

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 151

division of the genus does not in any way indicate that this is primitive; it has
rather been produced independently in the two groups, by the action of the
same three fundamental factors.

The phylogenetic tree thus shows two main stems extending in different
directions from a common origin. By the application of the three general
principles directing the evolution of the pattern of the ground color, it is possible

seta

fitkocclletis Tlat'larval

fvrphurosela ^^

; S

; ^

TUt -Ureal

MkoJleU^ 1

^roup.

Jypical ^

'\ ^

/'

. LreTnastohomhycia

Cl-ncestral /orm,

Fig. 25. Phylogenetic tree showing structural rela-
tionships of the groups of Lithocolletis .

Clncestral Tcrrw.

Fig. 26. Phylogenetic tree showing relative time of
origin of the groups of Lithocolletis.

to retrace the steps in the origin of the different groups of species. The evo-
lution of the species of the ''cylindrical-larval group/' which will be discussed
first, is illustrated on Plate III.

In all of the species except hataviella (Fig. 58), there has been a fusion
of Bands I and II along the costa.

If evolution in the shape of Bands II and III has been in the same direction
and at the same rate, and the outward growth of the middle of each has not
been sufficient to bring it in contact with the band following, and this outward
growth has been compensated for by an equal growth toward the base of the
extremities of Bands III and IV, species characterized by two entire fasciae will
be produced, the first at the basal fourth, the second at the middle of the wing.

Where evolution ceases early, all or most of the bands may be preserved as
transverse bands. L. tiliacella (Fig. 1) has been differentiated as a species
when, except for the fusion of I and II on the costa. Band IV alone had come
in contact with the band beyond. Specific differentiation, that is, the production
of dark margins, etc., has also halted soon, as we observe that no dark scales
are present on the last three bands. If development along the same line proceeds
farther, and by the farther outward growth of the middles of Bands II and III
and the extension inwardly of the extremities of III and IV, the fascise between
them become distinctly bent, the pattern of the ground color resembles that of
oregonensis. There has also been an accompanying gradual deepening of the
ground color. Specific differentiation at this level over a long period, as evi-
denced by the black apical spot, has given rise to oregonensis (Fig. 2). If,
on the other hand, evolution of the first three bands stops at the level of tiliacella
but continues in the apical four bands until IV, V, VI and VII are continuous
along the middle of the wing, species of the type of the European L. nicellii will
be produced.

152 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

If the bands continue to become more and more bent outwardly and the
fascise between them gradually narrowed, forms of the type of fragilella (Fig. 3)
will be produced. If now the action of the third law comes into play, the
outer edges of the bands on the margins will move toward the base; if the
extremities of the bands are extended toward the base at the same rate, very
acutely angled fascise will result, as in celtifoliella (Fig. 4) .

If the extension of the ground color toward the base in Band VI has taken
place rapidly, there is an entire obliteration of white between V and VI, which
then form a single uniformly colored band. In tritcenianella (Fig. 5), affinis
(Fig. 6) and ostensackenella (Fig. 7), this has occurred, Bands II, III and
IV remaining nearly straight. In mariceella (Fig. 8), these bands have
become bent outwardly. In ostensackenella, while the bands have remained
straight, there has been a very decided advance and specialization in other
directions; namely, in the structural modification of the scales forming the
fascise and spots and in the darkening of the base of the wing into a deep brown
except near the base of the dorsum. There has also been considerable modi-
fication in the larva, which has become more flattened and in which the sides
of the segments project more prominently than usual. It also deviates from
the typical habits of the genus in that it leaves the mine to pupate.

We may regard Porphyrosela as having originated from an ancestor with
this type of marking. The recent origin of the species within it (Fig. 9) is
indicated by the deep ground color, dark margins on either side of the white
fascise and the structural modification of the white and the purple scales.

If the extension distally of each of the bands along the cell has continued
until it has reached the band beyond and the inner extremities of the bands
have been produced basally at a slower rate, the result is the production of pairs
of opposite white streaks.

In auTonitens (Fig. 10), it is probable that the fascise were considerably
narrowed before these processes began, thus accounting for the almost per-
pendicular position of the white streaks.

In the remaining groups of species, these processes began early and the
pairs of streaks are more or less oblique. There is first produced a form in
which the ground color is blocked out as in argentinotella (Fig. 11). Occi-
tanica (Fig. 12) is derived from argentinotella by the obliteration of the first,
fourth and fifth costal streaks.

If now a movement of the extremities of each band toward the base is
accompanied by a corresponding extension of the extremities of the band beyond
it toward the base, very obhque white streaks result. Where these movements
have taken place at an approximately equal rate in all of the bands, forms such
as leucothorax (Fig. 13) will be produced. The outer edge of Band II on the
costa has been retracted completely to the base so that the first white costal
streak begins at the base. This species furnishes further evidence of the long
continuance of evolution by the well developed apical spot.

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 153

If on the dorsal margin^ the retraction of pigment from the outer edge of
Band III takes place more rapidly than the extension proximally of the extremity
of that band, its dorsal portion will disappear, and the second white dorsal
streak will coalesce with the first; fitchella (Fig. 14) typifies this result. The
extension of color onto a white streak has probably taken place most rapidly
just within the dorsal margin; the evidence for this conclusion rests upon the
observed concavity of the outer edges of the white streaks and their greater
width on the margin.

If the most rapid movement, as is usually the case, has been that of the outer
edge of Band II toward the base, the first result will be the union of the upper
inner edge of the first white dorsal streak with the short oblique streak at the
base separating Bands I and 11. A small streak of ground color will be left
between them on the dorsum; this soon becomes a small spot removed from
the dorsal margin by the more rapid shrinking away of Band II on the margin.
Salicifoliella (Fig. 15) and tremidoidiella (Fig. 16) have been differentiated as
species at this level.

The small spot of ground color just referred to becomes smaller and smaller
and finally disappears. The base of the dorsal margin is then occupied by a
white patch whose outer edge is the inner edge of Band III. Differentiation at
this stage has given rise to hageni (Fig. 17), arhutusella (Fig. 18) and insignis
(one variety) (Fig. 19). In hageni and insignis, the first white costal streak
has disappeared entirely. Following closely behind the retraction of pigment
from the outer edge of Band II on the costa, the extremity of Band III has
been produced along the costa to the base; phylogenetically then, this is the
second time pigment has appeared along the base of the costa. In arhutusella
(Fig. 17) the first costal streak has been involved in a change similar to
that of the first dorsal with which it has united. In these three species equally
distinct dark margins have appeared on either side of the white streaks, and
there has been a very decided darkening of the scales of the ground color between
the streaks, indicating that after the general phylogenetic evolution of the
pattern of the ground color had come to a standstill, specific differentiation pro-
ceeded rapidly and for a considerable period of time. In the other variety of
insignis (Fig. 20), Band III in its dorsal half has been produced to the
base, most rapidly just above the dorsum, leaving a white median basal streak
and a white spot on the dorsum near the base. This then is the method by which
a median long basal streak has originated; its basal part is therefore homologous
with the white streak separating Bands I and II; its apical half is homologous
with a part of the white streak separating the dorsal portions of Bands II and III.

Where the extension of Band III to the base has followed immediately on
the progressive withdrawal of pigment from the outer edge of Band II, without
the intervening cessation in development found in insignis, the configuration of
the color areas is that of the group of species of which obscuricostella (Fig. 25)
and cratcegella (Fig. 36) are typical examples. In most of the species of this

154 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

group, there has also been a retraction of pigment from the outer edge of
Band III, and at the same time, the extremities of Band IV have been produced
toward the base so that the first pair of white streaks are very obhque and are
situated before the tip of vein 16 which in more primitive forms determines the
position of these white marks. The obhteration of the first white costal streak
has not been attained in all the forms which reached this level before evolution
in the pattern of the ground color halted and species differentiation began,
since it is preserved in some species and lost in others.

Within this group the development of the dark markings in the apex has
taken two different directions; the black marking in the apex may take the form
of a round apical dot, or of a streak or patch of black scales of variable shape.
Of the species with apical dots, carycealhella (Fig. 21) lags somewhat behind the
other species; the suffusion of the base of the wing with pale yellow is not
yet complete, but the first costal streak has been lost. Development has been
more rapid toward the costa than in the dorsal half of the wing; the converse is
true for rileyella (Fig. 22) where the first white costal streak remains. In
conformity with the principle that within a group a darker ground color and the
presence of dark markings indicate a higher phylogenetic position, the sequence
of the remaining species with an apical dot will be ostrycefolieUa (Fig. 23),
olivceformis (Fig. 24), ohscuricostella (Fig. 25) and kearfottella (Fig. 26). In the
last two species all trace of a white spot near the base of the dorsum has
vanished.

Of the species characterized by more or less irregular patches of black scales
in the apex, populiella (Fig. 27) is probably the most primitive from the point
of view both of color areas and markings. Among several of the species,
namely, diaphanella (Fig. 28), salicivorella (Fig. 29), deceptusella (Fig. 30), alni
(Fig. 31) and alnicolella (Fig. 32), the pair of streaks in the middle of the wing
are not very oblique, that is, there has been a more restricted action of the last
two of the three processes of evolution than is observed in the five following
species. In ledella (Fig. 33), incanella (Fig. 34), scudderella (Fig. 35), cratcegella
(Fig. 36), malimalifoliella (Fig. 37) and propinquinella (Fig. 38), these streaks
are very oblique.

If evolution of the pattern of the ground color, instead of halting at the
level of the preceding group of species, proceeds farther, it takes place in two
directions.

1. The base of the wing may be uniformly suffused with ground color, so
that no white streaks remain when evolution ceases, and therefore no dark
markings develop. The configuration of the ground color in the remainder of
the wing has remained constant during these changes. This is the condition of
affairs in sexnotella (Fig. 39), ceriferella (Fig. 40) and obsoleta (Fig. 41). In ohsoleta
the pattern remained fixed for a short time, and the development of dark pig-
ment in the scales adjacent to the white streaks was initiated, then the exten-
sion of all the bands toward the base was resumed, checking further develop-

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 155

ment of dark pigment. The result is a tendency toward a uniform distribution
of ground color over the entire wing.

It may be objected that the color pattern of the three species just discussed
could have been derived directly and in a more simple manner from that of
L. argentinotella (Fig. 11) by the broadening of the extremities of Bands
II and III toward the base, resulting in the obliteration of the white streaks
separating Band I from II and Band II from III. On this hypothesis the
greater changes have occurred in the middle of the wing since the oblique pair
of streaks in the middle of the wing have been produced by the combined action
of the second and third processes of evolution. This is not in agreement with the
facts observed up to this stage in the evolution of the color pattern, namely,
that the greatest changes have occurred near the base of the wing, nor with the
general principle that development is more rapid in the more proximal parts
of an organism.

2. The shrinking away of the ground color at the extremities of the bands
may continue and the outer edges of Band III may be involved to a greater
extent than hitherto observed. In albanotella (Fig. 42) evolution in this
direction has ceased comparatively early, the darker ground color and heavier
dark margins indicating the longer time occupied in specific differentiation.
In argentifimhriella (Fig. 43) the withdrawal of color from the base of
the wing has been almost complete and there has been no later extension of
color toward the base from Band IV. In lucidicostella (Fig. 44) there is a
slight extension of color along the dorsal margin; the halting in development has
been too recent for a dark line of scales to develop on the streak of ground color
extending along the upper side of the cell. In clemensella (Fig. 45) this
streak has also disappeared. In the four preceding species, evolution has taken
place more rapidly on the dorsal than on the costal margin of the wing. In
quercialbella (Fig. 46) and trinotella (Fig. 47) the costal half of Band III
has also shrunk away; the only transverse bands then remaining are the
four apical ones. In the European fauna, the final consummation of this process
is witnessed in such species as L. cramerella and L. tenella. Development has
halted at different levels, as in our species, and dark streaks have appeared at
the edges of the bands. Gradually then bands more and more distal have
shrunk away, leaving behind them the dark fuscous streaks on an almost white
ground.

In all of the groups whose evolution has been traced, modification in the
shape of Bands II and III, in the early phases of evolution has progressed in the
same direction and at approximately equal rates, so that the bands are of
similar shape (cf. argentinotella, Fig. 11).

In the following species evolution has proceeded independently in each of
these bands.

Changes may proceed very slowly on the inner edge of Band IV and outer
edge of Band III; the tendency is then toward the preservation of a median

156 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

white fascia. If the outward growth of Band II along the cell is comparatively
rapid, a pair of opposite streaks will be produced at the basal fourth. These
changes have taken place very early in the phylogenetic history, and if evolution
of the pattern of the ground color ceases here, species of the type of morrisella
(Fig. 48), uhlerella (Fig. 49) and rohiniella (Fig. 50) will develop. The
laws governing specific differentiation have had an opportunity to operate
over a long period of time, and very highly specialized individual characters
have developed. In rohiniella evolution of the pattern did not cease quite as
soon as in the other two species, the fascia has been narrowly separated into
two streaks and all of the streaks have become more oblique. In consequence
of the later beginning of specific differentiation, the individual characters are
not as highly specialized as in uhlerella and morrisella.

If the edges of the first pair of streaks move toward the base, and there is
no corresponding extension of the extremities of Band III toward the base,
species of the type of lucetiella (Fig. 51) will be produced.

Where Band III has been extended to the base, there is a basal streak very
similar to that of cratcegella (Fig. 36) and the other species of that group.
Martiella (Fig. 52), gemmea (Fig. 53) and symphoricarpella (Fig. 54) have
originated thus. In symphoricarpella the basal streak is very small — ^the ground
color has shortened and narrowed it.

If the outer edge of Band III instead of remaining straight has grown distally
along the cell and at the same time the movement of its extremities toward the
base has been accompanied by a compensating extension of the extremities of
Band IV toward the base, angulated fasciae will result; apicinigrella (Figs.
55a, 556) and celtisella (Fig. 56) have been thus differentiated. Finally
this angulated fascia may become very acutely angled through the shrinking
away of Band III and the white may be extended along the margins to the base
of the wing. At this stage a very slight outward growth of the middle of Band III
will divide the fascia into very oblique streaks; basistrigella (Fig. 57) has
been thus produced. In this species the separation of a white fascia into a
pair of streaks was the last step in evolution; in the other species characterized
by opposite streaks, this process has been one of the earliest events in evolution.

Bataviella (Fig. 58) is an early isolated offshoot from the main stem;
it alone of all the species has preserved, in some of its varieties, Band I at the
base of the wing distinctly separate from II. Specific differentiation has ex-
tended over a length of time sufficient for all of the scales of the ground color
to become dark tipped.

The second main stem of the phylogenetic tree, which has given rise to
Cremastohomhycia and the ^^flat-larval group," is illustrated on Plate IV.

The species of Cremastohomhycia and of the ^^flat-larval group'' have fol-
lowed very similar lines of development. In Cremastohomhycia the result has
been modifled somewhat by the shorter and very oblique, almost horizontal
position of the white streak separating Bands I and II. A white basal streak

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 157

when formed is therefore nearer the dorsal margin. The chief forces acting in
the evolution of the pattern in this subgenus are the more rapid growth distad
of Band II as compared with Band III and the extension proximad of the ex-
tremities of Bands III and IV. The sequence of the species is as follows:
grindeliella (Figs. 59a, 5%), solidaginis (Fig. 60), ambrosiella (Fig. 61),
ignota (Fig. 62), verbesinella (Fig. 63). Grindeliella shows the basal streak in
actual process of formation; where the dorsal arm of the first fascia is wanting,
the black scales at the tip of the basal streak are homologous with those
on the outer margin of the fascia. All show the tendency which arises inde-
pendently throughout the ^^flat-larval group'' also, toward the complete fusion
of Bands V and VI. None of these species can be regarded as very primitive.
The similarity to a group of species in the division of the genus shortly to be
discussed is due to the fact that the same laws underlie their evolution.

A tendency similar to that witnessed among several of the species of the
first division of the genus, toward the preservation of Bands II and III in a shape
approaching the primitive, prevails among many of the species of the second
division of the genus. Obstrictella (Fig. 64) has preserved Bands II, III
and IV in a very primitive condition; during its differentiation, however, the
white streak at the base has been entirely obliterated.

Above this level, evolution takes place in four divergent directions, as repre-
sented on Plate IV.

In one of these branches. Bands II, III and IV tend to remain separate.
Within this group, the first step in evolution is the outward extension of the
middle of Band IV to meet Band V; at the same time the bands become displaced
very slightly toward the base of the dorsum; the fasciae between them are not
as yet angulated. Specific differentiation at this level has produced tubiferella
(Fig. 65), in which all white markings beyond the fasciae, except sometimes
a minute white dot before the apex, have been suffused with ground color;
aceriella (Fig. 66) and hamameliella (Fig. 67), the latter more advanced
because of its deeper ground color; ostryarella (Fig. 68), corylisella (Fig. 69),
cesculisella (Fig. 70) and guttifinitella (Fig. 71). The loss of the white
streak at the base seems to be accomplished after the principal changes
in the configuration of the ground color have been brought about, that is,
after specific differentiation has set in. The fact that the white streak
extends entirely across the wing, separating Bands I and II completely in
such species as hamadryadella (Fig. 77) and umbellularice (Fig. 78) which are
by no means primitive in regard to the shape of the areas of ground color, and
observations on the development of color in the pupal wings of cesculisella,
support this hypothesis.

In carycefoliella (Fig. 72) and lentella (Fig. 73), the bands have become
quite distinctly though obtusely angulated.

Above this level evolution has taken divergent paths. In one direction,
evolution in the shape of Bands II, III and IV has progressed in the same manner

158 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

and at the same rate, so that although the fascise between them have become
acutely angled, they have remained parallel. Macrocarpella (Fig. 74) and
cmcinnatiella (Fig. 75) have been differentiated at this level. These two
species are characterized by an unusual number of dark tipped scales; the
fascise are also internally dark margined near the costa. In saccharella (Fig.
76) evolution has continued farther, and other changes of a different char-
acter have resulted. There has been a slight displacement of the dorsal obHque
streaks toward the base; the first of these is brought into contact with the basal
streak. The costal portion of the first fascia has been replaced by the ground
color.

In the other direction, where evolution has progressed at a more rapid rate
on the outer edge of Band III and inner edge of Band IV, than on the outer and
inner edges of Bands II and III respectively, the second fascia will be more
acutely angled and its arms more obhque than the first fascia. Such is the
condition in hamadryadella (Fig. 77), umbellularice (Fig. 78) and agrifoliella
(Fig. 79). In the first of these there is but little divergence in direction
of the two fascise, evolution in this direction having halted compara-
tively early. Specific differentiation progressed rapidly and dark pigment
developed along the outer as well as the inner edges of the bands. At this stage
regressive changes have commenced; the bands have shrunk away toward the
base, most rapidly at the extremities, but also in the middle, leaving behind them
the Hues of dark scales which marked their former outer edges. In umbellularice
and agrifoliella, progressive evolution continued longer, and the arms of the
second fascia are very oblique compared with those of the first. Specific differ-
entiation has continued longer in agrifoliella than in umbellularice; the recent
disappearance of the white fascia separating Bands I and II is shown by the
presence of black scales external to its normal position and by the occasional
presence of a few pale scales anterior to them.

In all of these species characterized by two white fascise, there is a tendency
toward a disproportionate acceleration of the outward extension of the bands
along the cell, with the attendant result that the fascise are often interrupted at
the angle.

In a collateral line of development, there has been an early acceleration in
the processes of evolution affecting the outer edge of Band II and the inner
edge of Band III. As a result Band II becomes fused with Band III over a
considerable space in the middle of the wing, at a period when III and IV are
still separated from one another by an angulated fascia. Fletcherella (Fig. 80),
arcuella (Fig. 81), betulivora (Fig. 82), australisella (Fig. 83), bethunella (Fig. 84),
chamber sella (Fig. 85), cervina (Fig. 86), platanoidiella (Fig. 87), castaneceella
(Fig. 88) and fasciella (Fig. 89) have followed this path of evolution. Chamber-
sella (Fig. 85), cervina (Fig. 86) and platanoidiella (Fig. 87) have been differ-
entiated subsequent to the first five species; fusion of Bands II and III has been
complete in the dorsal half of the wing.

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 159

In castanemella (Fig. 88) and fasciella (Fig. 89), evolution has progressed
more slowly on the outer edge of Band III and inner edge of Band IV; the
median fascia remains almost straight, but the other white markings have
almost disappeared.

In two species, nemoris (Fig. 90) and gauUheriella (Fig. 91), the relations
between the progressive changes in Bands II and III have been reversed. The
tendency here is toward a fusion of Bands III and IV earlier than II and III,
with the result that the second fascia is separated into opposite spots earher than
the first. GauUheriella is the more advanced of these two species.

In all of the groups whose evolution has just been traced, the action of the
third process of evolution, namely, the retraction of pigment from the outer
edges of the extremities of bands, has been of minor importance. In the evo-
lution of four species, mediodor sella (Fig. 92), quercivorella (Fig. 93), conglomer-
atella (Fig. 94) and ulmella (Fig. 95), the characteristic longitudinal white streak
has originated through the action of this process at the extreme dorsal end of
each of Bands II, III and IV in sequence, followed by the extension of color
toward the base from Bands III and IV, just within the dorsal margin. This is
analagous to the process which has occurred in species of the first division of
the genus, where it was observed that the most rapid extension of a band into
the white streak preceding it took place just within the margin. These processes
occur first near the base, then the bands more and more distad are involved.
In mediodor sella (Fig. 92) Band IV has not yet been produced toward the base
to obliterate the dorsal arm of the median fascia; in quercivorella (Fig. 93) this
advance has been made. In conglomeratella (Fig. 94) and ulmella (Fig. 95), the
regressive changes have been extended to Band IV, which has shrunk away on the
margin, allowing the white obhque streak to move proximally along the margin
until it has united with the more proximal portion of the longitudinal streak.
This theory of the origin of the longitudinal dorso-basal white streak is sub-
stantiated by the variations in its width. It is widest at points correspond-
ing to the original extremities of the white fasciae, and narrowest immediately
behind them.

From the foregoing account of the evolution of the groups and species of
Lithocolletis, it is evident that the general evolution of the pattern of the ground
color may halt at any level and the species differentiated at this level constitute
a natural group of species more closely related to one another than to any other
species. The level at which evolution has come to a standstill is no absolute
criterion of the age of the individual species within a group. A species with but
few specialized specific characters in a group whose color pattern marks it as of
recent origin may be older phylogenetically than one of a more primitive group,
where the high degree of specific differentiation attained marks it as of recent
origin. It is to be expected that the most highly differentiated species, that is,
those where the darker colors predominate both as regards ground color and
markings, would be found in groups where the pattern of the ground color has a

160 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

less degree of modification; the earlier evolution in this direction has halted, the
longer will have been the time available for specific differentiation. On the
other hand, species belonging to groups recently estabhshed and hence of high
phylogenetic position, may be almost lacking in the darker colors. Actual
examples of each of these possibilities are furnished among our species; morrisella
(Fig. 48, PL III) and fragilella (Fig. 3, PL III) are examples of the first, quercial-
bella (Fig. 46, PL III) of the second possibility.

III. GENERAL DISCUSSION.

Ontogenetic and phylogenetic studies upon the development of the color
pattern in LithocoUetis lead to the conclusion that evolution has taken place in
definite directions under the action of a few definite laws; in other words, that it
has been orthogenetic. It was shown that the primitive color pattern, the first
to be laid down upon the white wing, is a series of seven uniformly colored, pale
yellow transverse bands. From these, the color areas, which form the ground
color of the various groups of species, have been derived. The first tendency
observed is a uniform widening of the bands. Evolution in the shape and extent
of the primitive bands of ground color, as has been shown (see page 141), has been
brought about by three definite processes, two of which were observed in actual
operation in the development of color in the pupal wings; the third is a necessary
inference from the lack of agreement between the primitive positions of the bands
and their positions in the more advanced species. The most far-reaching and
widespread changes have taken place toward the base of the wing proximal to
the transverse vein.

The final result of the action of these three processes is a uniform ground
color; this uniformity will be attained first near the base of the wing, since it is
here that evolution has proceeded most rapidly. The color bands in the apical
half of the wing, which were laid down later, are modified at a slower rate and
retain more nearly their primitive shape. These observations are in agreement
with the general principle that development is more rapid in the anterior and
proximal parts of an organism.

The observed evolution in the pattern of the ground color suggests that
the uniform yellowish ground color which suffuses the wing in the higher Lepidoptera,
beginning at the base and spreading distalward, is the outcome of a phylogenetically
older type of marking, originally banded, and later fused to a uniform color, and
that the markings are a second series superimposed upon the first.

If the disposal of the transverse primitive bands of ground color is de-
pendent upon the positions of the nervures, as my observations on LithocoUetis
have indicated, then in the primitive color pattern of the ancestral Lepidoptera
where more veins are present and their arrangement on the margins is more
symmetrical, the number of these bands should be considerably augmented.
As evolution has proceeded, this primitive type of color pattern would have been
modified in various ways. Attendant upon the degradation in venation in some

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 161

of the lower groups, there must have been immediate and complete fusion of
bands occupying adjacent interspaces, when the vein which separated them was
lost. Modification in the shape of the wings, resulting in changes in the relative
positions of the nervures, must also have been an important factor in producing
correlated changes in the primitive color pattern by bringing about displacement
and fusion of some of the bands. Just what modifications in the color pattern
would have been produced by the direct action of these conditions, could only
be determined by a complete knowledge of the actual series of changes which
have taken place in the wings of Lepidoptera. Beyond, and apparently inde-
pendent of these factors, there is the evolution in the shape and extent of the
color areas themselves.

It was found that dark markings, properly so-called, appeared at the limits
between ground color and unpigmented areas. Increase in breadth of a dark
marking takes place away from an unpigmented area. The dark markings do
not appear until the edge of a color area has remained constant for an appreci-
able length of time. Since they appear at the edges of color areas, their shape
and position are in great part determined by the same laws which control the
evolution of the pattern of the ground color. Markings once formed tend to
become permanent and immovable. Thus we find that the suffusion of an un-
pigmented area with ground color subsequent to the formation of a dark streak
does not affect its permanency; the shrinking away of ground color leaves a dark
streak or line isolated in an otherwise unpigmented area. Therefore, although,
during the early stages of evolution, the development of markings was entirely de-
pendent on the configuration of the ground color, these markings once permanently
established in the race, tend to reappear independently of the ground color. When
this level has been reached in racial history, we have a series of transverse markings
appearing on a uniform ground color. This is the condition in all of the higher
Lepidoptera. In the ontogeny of such groups, there is no evidence to show that
these markings are a second series superimposed upon the earlier primitive one.

If the theory of the origin of dark markings on opposite edges of primitive
bands of ground color is a valid one, we should expect to find in the higher types,
that there is a tendency for dark stripes to recur in pairs, and that fusions at any
single period would be more apt to occur between any two consecutive stripes
than between three consecutive stripes. Observations confirm this hypothesis.
Among many of the Pyralids, the dark markings are formed of two dark lines
connected by a paler color; each of these marks is separated from a similar pair
of lines by ground color. Even in the butterflies, the tendency toward grouping
of dark marks in pairs is witnessed. Elmer's ('97) series of adult forms of Papilio
and von Linden's ('98, '02) figures of the pupal development illustrate this
principle. In the pupal development of the wings of Papilio machaon, the
tendency for the dark scales to appear first at the edges of bands was noted;
bands thus formed have, on my hypothesis, originated from the fusion of dark
stripes situated on either side of a primitive band.

162 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

The great diversity to be observed in the position^ shape and extent of the
dark markings in the various groups of Lepidoptera, is due to differences in the
direction of evolution of the pattern of the ground color, and the different levels
at which it has halted, permitting the dark markings to develop.

The actual appearance of dark pigment in certain defined positions is prob-
ably due to physiological factors; just why these processes should occur con-
tiguous to unpigmented areas is largely a matter of conjecture. It is possible
that the chemical changes involved are oxidative in their nature; the presence
of air in the unpigmented hollow scales may possibly accelerate chemical proc-
esses along this line. This supposition will however not account for the phylo-
genetic permanency of such markings after the ground color covers the wing
uniformly.

The investigations of previous workers upon the color patterns of Lepi-
doptera have been limited to the markings, that is, to that series which I have
designated as secondary and superimposed upon an original primary color
pattern. In as much, however, as these markings have originally been deter-
mined by the configuration of the pattern of the ground color, the laws con-
cerning their places of appearance should agree with my observations on the
evolution of the pattern of the ground color; it is on the contrary not to be
expected that these laws concerning behavior of the markings will apply in their
entirety to the comparatively primitive and phylogenetically older type of
markings in this genus. The laws governing the color pattern have been based
upon more highly developed markings often of comparatively recent origin
phylogenetically. For convenience in comparison a statement of some of these
laws of coloration follows. The following are those advanced by Mayer ('97) :

^^(a) Any spot found upon the wings of a moth or butterfly tends to be
bilaterally symmetrical, both as regards form and color, the axis of symmetry
being a line passing through the center of the interspace in which the spot is
found, and parallel to the direction of the longitudinal nervures. (b) Spots
tend to appear not in one interspace only, but as a row occupying homologous
places in successive interspaces. Indeed we almost always find spots arranged
in linear series, each similar in shape and color to the others and occupying the
center of its interspace, (c) It is interesting to notice that bands of color are
often made by the fusion of a row of adjacent spots; and conversely, chains of
spots are often formed by the breaking up of bands, leaving a row of spots
occupying the interspaces, {d) The most common method of disappearance
is a shrinking away of the band at one end. It is very common to find bands
shrinking away at one end. This is a special case of Bateson's law that the ends
of a linear series are more variable than the middle. Sometimes however they
shrink away at both ends, and very often, they break up into a row of spots,
which may then contract into the centers of their interspaces and finally dis-
appear. It is worthy of note that it is very rare to find a band breaking at the
middle of its length and each half receding from the other, (e) The position of

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 163

spots which are situated near the edge of the wing is largely controlled by the
wing-folds or creases."

Eimer has derived the various types of coloration from a series of transverse
stripes (termed by him longitudinal because parallel to the longitudinal axis
of the body) which break up into spots which fuse crosswise to form markings
parallel to the longitudinal nervures, and finahy fuse to a uniform color. New
markings appear on the body of the animal from behind forwards and from above
downwards, or conversely, whilst the old ones disappear in the same direction
and succession.

I have shown that the ground color has been derived from a series of such
transverse elements; conversion into a uniform color has taken place directly
by an extension of color onto previously unpigmented areas, distad through the
middle of the wing in the ceh, proximad on the margins. The positions of the
markings are arbitrarily determined. They are dependent upon the level at
which evolution in the pattern of the ground color acting in a few definite direc-
tions has come to a standstill. The actual appearance of the markings is prob-
ably due to physiological activities whose nature is not understood. The first
markings are usually transverse, rarely longitudinal when the modification
in the ground color has been far-reaching. These transverse markings are in
the nature of narrow lines or streaks, phylogenetically the forerunners of the
bands of higher Lepidoptera. Since the streaks and lines of dark scales in Litho-
colletis still retain their primitive condition as originally laid down, it is not to he
expected that the laws which Mayer has given, based upon the transformation of
spots and bands in the higher Lepidoptera, would find wide application in this
group. Portions of a pair of such streaks in Lithocolletis will be found at the
margins and in each interspace which the band from which they are derived
crosses. Breaking up of bands formed by the growing together of these streaks
will result in conditions such as Mayer has described for spots and series of spots.
Any spot such as the apical spot is bilaterally symmetrical and occupies the
center of its interspace. The primitive bands from which the ground color has
been derived and which have been regarded as a primary set of markings upon
which a second set, the markings proper, have been superimposed, have exhibited
the tendency toward shrinking away at each end.

The ultimate causes which determine the definitive positions of the primitive
bands are unknown. Von Linden has found that in the lower orders of insects,
the coloring matter has a tendency to collect on transverse veins; this however
does not bring us any nearer to a real explanation of the phenomenon of colora-
tion. The probability that the origin of the markings, which appear secondarily,
may be traced directly to physiological factors, renders it reasonable that the
appearance of the primary series is conditioned by like physiological and morpho-
logical factors within the organism itself and is independent of external factors.

Evolution in the pattern of the primitive series of bands has been shown to
have taken place in a few definite directions; the diverse patterns which have

164 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

been produced are the result of different combinations of three general processes
already discussed.

IV. SUMMARY.

Observations on the development of color in the pupal wings and a com-
parison of the color pattern in adult forms of Lithocolletis have shown that in
his group the primitive color pattern is a series of seven uniformly colored pale^
yellow transverse bands, separated from one another by unpigmented areas.
The disposal of these bands is dependent upon the course of the longitudinal
nervures, since the points of origin or the tips of the veins mark the positions of
the unpigmented fasciae between the bands.

From this primitive color pattern, the several different types of color pattern
in this genus have been derived. Evolution has taken place in definite direc-
tions, under the action of three general processes, which were found to be suf-
ficient to explain the origin of the different color patterns. These processes have
been stated thus: (a) The middle portion of a band may be produced distally
until it comes in contact with the band beyond it. (6) The extremities of a
band may be broadened by being produced proximally. (c) The extremities
of a band may be narrowed by the retraction of pigment from their outer edges.
The particular result in any case, that is, the configuration of the color areas,
is due to the manner and extent of the action of these processes.

During pupal development, these phylogenetic changes are repeated only
to a very limited extent; recapitulation is confined to those species or to those
portions of the wing in which there has been the least modification of the primi-
tive transversely banded type of marking.

These bands either in their primitive or modified shape, constitute the
ground color. Upon this ground color, a second darker series of elements, the
markings proper, also usually transverse, are superimposed. The different levels
at which evolution in the pattern of the ground color has halted, that is, the
configuration of the areas of ground color, are the important factors in deter-
mining the phylogenetic sequence of large groups; the particular colors and the
markings determine the positions of the species within those groups.

These markings appear at the limits between ground color and unpigmented
areas, as one or more lines of dark scales along the edge of a band adjacent to a
white fascia or streak. Their appearance then is probably due to physiological
factors within the organism itself, and is independent of external conditions.
In the ontogeny, the relative time of appearance of the dark margin of any band
is dependent upon the time when the edge of that band became fixed in the
phylogeny. Before dark margins can develop, the edge of a band must have
remained fixed for an appreciable time. Hence, where bands have been laid
down late or have reached their present configuration late in racial development,
the dark margin will develop at a correspondingly late period in ontogenetic
development.

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 165

This principle holds good irrespective of whether or not there is a recapitu-
lation of phylogenetic changes in the configuration of the color areas during
pupal development. Therefore in the second and more recent set of characters,
there is an actual recapitulation. A dark marking once permanently established
in the race tends to reappear independently of the ground color, so that later
suffusion of the unpigmented area contiguous to it with ground color, or the
shrinking away of the ground color, does not affect its permanency.

Later in phylogenetic history, additional dark markings, other than those
contiguous to unpigmented areas, may appear. The development of these
characters in the pupa becomes much abridged, and concomitant with this,
their time of appearance is pushed back into the earlier stages of pupal develop-
ment, so that they may appear simultaneously with or even earlier than char-
acters which are much older phylogenetically. However, only those characters
permanently established and of long standing in a species exhibit this precocity
of development; recently acquired or variable characters appear ontogenetically
in the order of their phylogenetic sequence. Where certain characters appear
unduly early in pupal development, physiological factors probably act directly
in bringing about this result.

From a study of the phylogeny of the various groups of species, through
the application of the processes of evolution already enumerated, it was found
that the most far-reaching and widespread changes have taken place toward
the base of the wing, proximal to the transverse vein. The final result is the
production of a uniform ground color which will be attained earliest near the
base of the wing where evolution has proceeded most rapidly.

This observed evolution in the pattern of the ground color suggests that
the uniform yellowish ground color which suffuses the wing in the higher Lepi-
doptera, beginning at the base and spreading distalward, is the outcome of a
phylogenetically older type of marking, originally banded, and later fused to a
uniform color, and that the markings are a second series superimposed upon
the first.

The observations made clearly point to the conclusion that the evolution
of the color pattern in Lithocolletis has been orthogenetic.

166 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS.

LITERATURE CITED.

BeMMELEN, J. F. VAN

1889. Ueber die Entwickelung der Fiirben und Adern auf den Schmetterlingsflligeln. Tijdsch. der
Nederland. Dierkundige Vereeniging, Ser. 2, Deel 2, pp. 235-247.

Braun, Annette F.

1908. Revision of the North American Species of the Genus Lithocolletis Hiibner. Trans. Am. Ent.

Soc, vol. XXXIV, pp. 269-357, 5 pis.

1909. Phylogeny of the Lithocolletid Group. The Canadian Entomologist, vol. XLI, pp. 419-423.

Chapman^ T. A.

1902. The Classification of Gracilaria and Allied Genera. The Entomologist, vol. XXXV, pp. 81-88,
138-142, 159-164.

DlMMOCK^ G.

1883. Scales of Coleoptera. Psyche, vol. IV, pp. 1-11, 23-27, 43-47, 63-71

EiMER, G. H. T.

1897. On Orthogenesis and the Impotence of N^atural Selection in Species-Formation. An address

delivered at the Leyden Congress of Zoologists, Sept. 19, 1895. 55 pp. The Open Court
Pubhshing Co., Chicago, 1898.

Haase, E.

1893. Untersuchungen iiber die Mimicry auf Grundlage eines natiirhchen Systems der Papilioniden.

Bibliotheca Zoologica, Heft 8, Theil 1, 120 pp., 6 Taf.
1893a. Untersuchungen liber die Mimicry auf Grundlage eines natlirlichen Systems der Papihoniden.

Bib. Zool., Heft 8, Th. 2, 161 pp., 8 Taf. Enghsh translation by C. M. Child. Stuttgart,

1896, 154 pp., 8 pis.

Linden, M. von

1898. Untersuchungen iiber die Entwickelung der Zeichnung des Schmetterlingsfliigeln in der Puppe.

Zeitschr. fiir Wiss. Zool., Bd. LXV, pp. 1-49, pis. I-III.
1902. Le dessin des ailes des Lepidopteres, reserches sur son evolution dans I'ontogenese at la phylo-
genese des especes, son origine et sa valeur systematique. Annales des Sci. Nat., vol. XIV,
pp. 1-196, pis. I-XX.

Mayer, A. G.

1896. The Development of the Win'g Scales and their Pigment in Butterflies and Moths. Bull. Mus.

Comp. Zool, vol. XXIX, No. 5, pp. 209-235, 7 pis.

1897. On the Colors and Color Patterns of Moths and Butterflies. Proc. Bost. Soc. Nat. Hist., vol.

XXVII, No. 14, pp. 243-330, 10 pis.

Meyrick, E.

1912. Fam. Gracilariadae. Genera Insectorum, 128™*^ Fascicule.

SCHAFFER, C.

1889. Beitrage zur Histologic der Inseckten. Zoologische Jahrbiicher, Abth. flir Anatomic, Bd. Ill,
Heft 4.

Tower, W. L.

1906. An Investigation of Evolution in Chrysomelid Beetles of the Genus Leptinotarsa. Carnegie
Institution of Washington, Publication No. 48.

Urech, F.

1891. Beobachtungen liber die verschiedenen Schuppenfarben und die zeitHche Succession ihres
Auftretens (Farbenfelderung) auf den Puppenflugelchen von Vanessa urticse und V. oi.
Zool. Anz., Jahr. XIV, pp. 466-475.

EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS. 167

EXPLANATION OF PLATES III AND IV.

PLATE III.

Phylogenetic tree illustrating the evolution of the species of the "cylindrical-larval group
and of Porphyrosela. (The figures on this and the following plate are three times natural size.)

of Lithocolletis

Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.
Fig.

1.
2.
3.

4.
5.
6.

7.

8.

9.

10.

Fig. 11.
Fig. 12.
Fig. 13.
Fig. 14.
Fig. 15.
Fig. 16.
Fig. 17.
Fig. 18.
Fig. 19.
Fig. 20.
Fig. 21.
Fig. 22.
Fig. 23.
Fig. 24.
Fig. 25.
Fig. 26.
Fig. 27.
Fig. 28.
Fig. 29.
Fig. 30.
Fig. 31.
Fig. 32.
Fig. 33.
Fig. 34.
Fig. 35.
Fig. 36.
Fig. 37.
Fig. 38.
Fig. 39.
Fig. 40.
Fig. 41.
Fig. 42.
Fig. 43.
Fig. 44.
Fig. 45.
Fig. 46.
Fig. 47.
Fig. 48.
Fig. 49.
Fig,
Fig,
Fig. 52.
Fig. 53.
Fig. 54.
Fig. 55a.
Fig. 555.
Fig. 56.
Fig. 57.
Fig. 58.

50.
51.

Liihocollelis tiliacella Cham.

" oregonensis Wlsm. (not illustrated).

" fragilella F. and B.

" celtifoliella Cham.

" tritcenianella Cham.

" affinis F. and B. (not illustrated).

" ostensackenella Fitch.

" mariceella Cham.

Porphyrosela desmodieUa Clem.
Lithocolletis auronitens F. and B.

argentinotella Clem.

occitanica F. and B. (not illustrated).

leucothorax Wlsm.

fitchella Clem.

salicifoliella Clem.

tremMoidiella Braun.

hageni F. and B.

arhutusella Braun.

insignis Wlsm. (not illustrated).

insignis Wlsm.

carycealhella Cham.

rileyella Cham.

ostrycEfoliella Clem.

olivceformis Braun.

ohscuricostella Clem.

kearfottella Braun.

populiella Cham.

diaphaneUa F. and B.

salicivorella Braun.

deceptusella Cham.

alni Wlsm. (not illustrated).

alnicolella Wlsm.

ledella Wlsm.

incanella Wlsm.

scudderella F. and B.

cratcegella Clem.

malimalifoliella Braun.

propinquinella Braun.

sexnotella Cham.

aeriferella Clem.

obsoleta F. and B.

albanotella Cham.

argentifimhriella Clem.

lucidicostella Clem.

clemensella Cham.

quercialhella Fitch.

trinotella Braun.

morrisella Fitch.

uhlerella Fitch.

rohiniella Clem.

lucetiella Clem.

7nartieUa Braun.

gemmea F. and B.

symphoricarpella Cham.

apicinigrella Braun.

apicinigrella Braun.

celtisella Cham.

basistrigella Clem.

hatavietla Braun.

168 EVOLUTION OF COLOR PATTERN IN LITHOCOLLETIS,

PLATE

IV

Phylogenetic tree illustrating the evolution of the

spe

group '^ of Lithocolletis.

Fig. 59a. Cremastohombycia grindeliella Wlsm.

Fig. 595.

" grindeliella Wlsm.

Fig. 60.

" solidaginis F. and B.

Fig. 61.

" ambrosiella Cham.

Fig. 62.

" ignota F. and B.

Fig. 63.

" verbesinella Busck.

Fig. 64. Lithocolleiis ohstrictella Clem.

Fig. 65.

' tubiferella Clem.

Fig. 66.

' aceriella Clem.

Fig. 67.

' hamameliella Busck,

Fig. 68.

' ostryarella Cham.

Fig. 69.

' corylisella Cham.

Fig. 70.

' cesculisella Cham.

Fig. 71.

' guttifinitella Clem.

Fig. 72.

' carycefoliella Clem.

Fig. 73.

' lentella Braun.

Fig. 74.

' macrocarpella F. and B.

Fig. 75.

' cincinnatiella Cham.

Fig. 76.

' saccharella Braun.

Fig. 77.

' hamadryadella Clem.

Fig. 78.

' umbellularim Wlsm.

Fig. 79.

' agrifoliella Braun.

Fig. 80.

' fletcherella Braun.

Fig. 81. '

' arcuella Braun.

Fig. 82.

' betulivora Wlsm.

Fig. 83.

' australisella Cham.

Fig. 84.

' belhunella Cham.

Fig. 85.

' chambersella Wlsm.

Fig. 86.

' cervina Wlsm.

Fig. 87.

' platanoidiella Braun.

Fig. 88.

' castaneceella Cham.

Fig. 89.

' fasciella Wlsm.

Fig. 90.

' nemoris Wlsm.

Fig. 91.

' gaultheriella Wlsm.

Fig. 92.

' mediodor sella Braun.

Fig. 93.

' quercivorella Cham.

Fig. 94.

' conglomeratella Zell.

Fig. 95.

' ulmella Cham.

flat-larval

JOURN. ACAD. NAT. SCI. PHILA., 2ND SEP., VOL. XVI.

PLATE III

f *.

•,!f^-»~

hT

^*jg5^ff#

f .

alnv
31

\ 32

^'S*

^

^

W

m

-'' ^^^P >7f>>'» f^^#.

t""\: — ''

I ■'>

■^

1 'i 5/

28

1 ^^^

'f^ ^m€^'

\ f^

iG ,n

r:^^

' 14

13

. occitanica

nm

12

\

$p^

19

3)^

^

Jypical cl^lkocollet^i
Priimtivc Clncestor

:tK

oreqonensis
2

Seiren Unijormlij Colored Transverse Bands.

X,

A.KB. del.

^ y ^ Lremastoiaom/bivcia

ulcLi - 1 xrvaiUroap

m^

affini£

f^

BRAUN: Color Patterns in Lithocolletis.

JOURN. ACAD. NAT. SCI. PHILA., 2ND SER., VOL. XVI.

PLATE IV.

'K^

^ f*

M^

r-^' i

I ^^ t

:>!^-

^

^^;^

J Lat-larval Cjroup

«;>.*^^

Q^., -^

rr^^^

r

y^

few^

^

!?)«»'

"^ «

">»,

^

63

/^■StT^.

C r e 771 a s 1 boTTi bi^c I (2/

^ s:^r^-

rrimitive Cincestor

Ak

Oeven ilrLijormlij Lolored Jransverse JJand:
Jijpical jLitnocoiteti^

BRAUN: Color Patterns in LithocoHetis.